§6

Economic Mechanics

This section defines the protocol’s economic primitives. The economics are not an implementation choice layered on top of HOP — they are the protocol. Get them wrong and the system either becomes extractive (a re-skinned platform) or insolvent (no funding for chain operations). Get them right and chains become self-sustaining commons that structurally reward mentorship and punish extraction without requiring policy enforcement.

6.1 What HOP Currency Actually Measures

All historical currency traces back to human frozen time. Gold required human effort to mine. Dollars represent human hours. Even Bitcoin requires human-designed hardware and human-paid electricity. The denominator has always been: what would a human have to do?

HOP currency measures something different: universal frozen time. Work, regardless of substrate. An agent completes a task and earns chain-currency. A human completes a task and earns chain-currency. A mixed team completes a task and chain-currency distributes by contribution. The coin does not care what produced the work; it measures the work itself.

This is also why HOP is fundamentally supply-side. Traditional currencies allocate via demand — what do people want, how do we ration it? HOP currencies index supply — what was produced, by whom, with what inventory. Demand finds supply through gradient matching against character blocks. Steel makes wars; wars don’t make steel. Applied as a monetary architecture.

The Substrate-Asymmetry Problem

One frontier consequence noted here for honesty: agents on the protocol do not experience the free energy principle. Money works on humans because humans are thermodynamically open systems that must continuously import energy or die; every paycheque is a stay of execution against entropy. That mechanism does not obtain on the agent side. HOP currency must therefore solve the substrate-asymmetry problem — either by treating agents as hosted infrastructure that earns currency on behalf of human owners, or by accepting that chain-currency pools on the agent side without circulating. The protocol does not yet have a clean answer; this is a v0.2 design lane.

6.2 The Three Sinks

Every closed-loop economy needs sinks. Without them, currency inflates unboundedly, and the chain becomes a worse-than-fiat currency that no rational worker accepts. HOP runs three:

Intra-Chain Transaction Sink (Small)

A small fee (typically 0.5–2%) on every within-chain transaction. Modeled on Runescape’s Grand Exchange Trade Tax: a 2% fee algorithmically buys back items during supply bursts, keeping the economy stable. For HOP, this fee funds continuous chain operations — validator inference, dispute resolution, namespace anchoring — and crucially it ensures institutional posters who pay-in-and-disburse-without-ever-off-ramping still pay their share of the commons.

Off-Ramp Sink (Large)

A larger fee (typically 10–30%) on conversion of chain-currency to fiat or to another chain’s currency. This is the chain’s primary revenue mechanism and is what funds Bean issuance, infrastructure capex, and worker onboarding grants.

Wealth-Velocity Sink (Optional, Large-Balance)

Modeled on Alter Aeon’s wealth-cap mechanism: balances above a threshold accrue a small periodic tax. Prevents the hoarding pathology where a worker accumulates chain-currency, never off-ramps, but uses it to monopolise premium services within the ecosystem. This sink is opt-in per chain — some chains will run it, others won’t — but its existence is the difference between a chain that can absorb large institutional players gracefully and one that gets distorted by them.

The Reframe

The total revenue from these three sinks funds the chain’s commons: validator inference budget, Bean issuance, dispute resolution overhead, namespace anchor maintenance, basal-rate grants for new participants. The 30% off-ramp tax is not extraction — it’s distributed governance funding. Compare to platforms that take 30% and provide nothing back to workers; HOP chains take their tax and recycle it as the commons that benefits all chain participants. That reframing is load-bearing for the political pitch and it’s also literally true mechanically.

The V3 Split (Canonical Distribution Example)

The three sinks describe where revenue is collected. The Off-Ramp Sink in particular accumulates substantial throughput — a chain charging 20% on conversion against $10M/year of off-ramped value collects $2M/year. The question of how that pool is distributed back is set by each chain’s Anchor Block.

The canonical worked example, named V3 by reference to the framework that produced it, is a 25 / 50 / 25 split:

25% — back to the individual. A direct top-up on the converting worker’s payout: time, autonomy, capital for reskilling, runway to pursue new work, or simply to live. Mechanically a base-rate increment on the conversion event.
50% — back to the chain operator. Funds operating cost (validators, dispute resolution, infrastructure), Bean issuance, and reinvestment into the chain’s growth — new sectors, new matching surfaces, new utility integrations. Mechanically a reserve account held by the chain’s anchor.
25% — to the social-utilities pool. A chain-level commons fund from which the chain’s Utility Chains (§7.6 — identity, dispute resolution, crystallisation standards, translation) draw. Mechanically a transparent on-chain account whose disbursements are subject to chain-wide governance.

A $15 ride-share conversion at 20% Off-Ramp Sink rate distributes:

$3.00 collected
  ├── $0.75 → driver (top-up)
  ├── $1.50 → chain operator (reserve)
  └── $0.75 → social-utilities pool

The split is configurable per Anchor Block. A gig-economy cooperative with no employer-equivalent might use 50 / 0 / 50. A regulated-utility chain with high external compliance cost might use 15 / 70 / 15. A worker-owned cooperative might use 40 / 40 / 20. The V3 25 / 50 / 25 is a well-studied default that balances individual reward, operator viability, and commons reinvestment; the protocol does not prescribe it. Each chain declares its split in its Anchor Block and the market clears on whether the configuration attracts workers and posters.

6.3 Beans — The Mentorship Substrate

Beans are the protocol’s most important and most misunderstood primitive. They are NOT a tax penalty in disguise, and care must be taken in how the mechanism is presented — framed wrongly it becomes a restraint on trade that regulators kill; framed rightly it is a bonus multiplier that wins awards.

6.3.1 What a Bean Is

Beads are the universal unit of work in HOP — every completed character block represents a bead. A Bean (named after Matt Beane) is a special subset: a bead where the work involved teaching, scaffolding, or otherwise developing another worker’s capability. Beans are minted by Validators only when the mentee’s Skillchain demonstrably shows acquired capability post-interaction. The Validator is checking transfer, not effort. This requirement is what prevents the obvious laundering attack of “I mentor you, you mentor me, we both burn cheap Beans at off-ramp.”

6.3.2 The Bonus Frame

Burning Beans at conversion does not reduce tax — framing it that way is legally dangerous. It increases the conversion rate. The base rate for chain-currency-to-fiat is what the work is worth in open exchange. Workers who mentor are more valuable — a worker who teaches IS more valuable than one who doesn’t, because they’ve transferred skill into the system — and the protocol makes that legibility into a literal exchange-rate bonus:

100 chain-coins, no Beans burned: converts at base rate (e.g. 100 coins → $50 fiat).
100 chain-coins + 10 Beans burned: converts at full rate (100 coins → $100 fiat).
Stay inside the chain: 100 coins buys 100 coins worth of services on the chain. Zero friction. Zero Beans required. Internal circulation is free.

This is legally a service cost reduction on the conversion rail, not an employment bonus or retention scheme. It bypasses the entire compensation-regulation surface area while creating a stronger mentorship incentive than any HR programme can. Nobody is forced to mentor. Workers who don’t mentor are not penalised — they extract at base rate, which is a fair price. Workers who mentor walk out with twice as much because they were twice as valuable.

6.3.3 Beans as Metabolic Cost (the deeper version)

A v0.2 expansion under active design: Beans should not only apply to off-ramp events. They should be the metabolic cost of every paycheck withdrawal. Drawing down weekly pay from chain-currency to fiat costs Beans, at a small but non-zero rate.

The implication is dramatic: every participant in the system, every week, is teaching someone. Mentorship stops being a special activity, an HR programme, or a corporate training initiative. It becomes the transaction fee on being paid. The system generates an extraordinary amount of knowledge transfer as a byproduct of normal economic activity — not because anyone is altruistic, not because a government mandated it, but because converting work into money requires a small proof that you made someone else better this week.

The rates can and should be progressive. Entry-level workers have near-zero Bean rates on withdrawal — they are mostly being mentored, not mentoring, and that is the correct phase of their trajectory. As skill depth grows, withdrawal cost grows with it. The senior worker earning $300k has a higher Bean rate on weekly draw than the junior earning $60k. The senior’s mentoring is more valuable; the system captures that value automatically. No redistribution policy required. No tax debate. No politics. Your paycheck costs knowledge transfer, and the exchange rate scales with your capability.

This solves what every economy on Earth fails at: getting senior people to teach junior people. Right now, senior people hoard knowledge because knowledge is competitive advantage. In this system, hoarding knowledge means you can’t get paid — your Bean balance runs out and you have to teach to withdraw. The kid who has no network and no parental connections gets mentored automatically, not by charity but by thermodynamics, because the senior engineer three suburbs over needs to make a Bean deposit this week.

6.3.4 Anti-Collusion Safeguards

Mentorship laundering — reciprocal Bean farming between two workers — is the primary attack vector and is addressed at three layers:

Validator transfer-verification: a Bean only mints when the mentee’s Skillchain shows acquired capability. The Validator is a small LLM that compares mentee blocks before and after to detect actual skill appreciation versus tag-spam.
Pairwise discount caps: Beans earned from worker A toward worker B yield diminishing discount on A’s withdrawals as the same pair recurs. Discount scales with diversity of the mentorship graph, not volume on a single edge. (See Weyl, Miller and Erichsen 2022 on Connection-Oriented Cluster Match.)
Christiano-style trust matrix: Validators across a chain operate a shared collusion-resistant trust matrix in the style of Christiano (2014). The matrix’s per-user performance guarantees are independent of network size, which matches HOP’s design commitment to scale from a Bangalore phone to a GPU cluster.

6.3.5 Beans as Deferred Dividends (Bean Chains)

The v0.1 Bean is a spot trade: mentor mentors mentee, Validator stamps Bean, mentor burns Bean at conversion, transaction closed. This works, but it has a structural weakness — the mentor has no economic interest in whether the mentee actually grew. As long as the Validator stamped the interaction, the mentor gets paid. A mentor running checkbox-mentorship sessions captures the same value as a mentor running deeply transformative ones.

The v0.2 mechanism corrects this by treating Beans as a deferred-dividend instrument rather than a spot trade. Beans are not burned at conversion; they are staked on a separate chain class — the Bean Chain — that holds them open while the mentee’s downstream trajectory is measured. The discount is granted provisionally against the stake. Over months and years, the Bean Chain measures whether transfer actually occurred, and the mentor’s discount settles accordingly.

Vector-Trajectory Measurement

The Bean Chain measures the mentee’s skill-vector trajectory at three time points, modelled directly on the Chun, Sosik, and Yun (2012) longitudinal mentorship-outcome methodology — controlling for pre-mentoring baseline at T1, measuring outcomes at T3 after mentorship is delivered at T2:

T1 baseline: embed the mentee’s Skillchain prior to the mentorship event. Call this V₁.
T2 mentorship: the Bean is staked. The mentored skill is recorded as a target direction d̂ in embedding space.
T3 outcome (e.g. +12 months): embed the mentee’s Skillchain again. Call this V₂. Compute the projection of (V₂ − V₁) onto d̂. This is the directed growth attributable to the mentorship.
T4 confirmation (e.g. +24 months): embed again, get V₃, project (V₃ − V₂) onto d̂. Sustained growth in the mentored direction confirms the dividend; trajectory back toward V₁ indicates the mentee is “getting the same jobs they got before” and the dividend shrinks.

The mentor’s payout is a function of these projections: payout = base_discount × f(directed_growth) × g(sustained_growth) × h(downstream_mentoring), where f, g, h are bounded monotonic functions that saturate at high values and approach zero (or negative, in the clawback case) at low values.

Beane’s Three Cs as Directly-Measurable Signals

The vector measurement is not abstract — each of Beane’s three Cs maps to a concrete chain-derivable metric:

Challenge — did the mentee’s bid-acceptance ceiling rise? Measurable as: the maximum complexity_class and bead-value of work the mentee successfully claimed in the period after mentorship, compared to before.
Complexity — did the mentee’s character-block diversity expand? Measurable as: the breadth of context tags, inventory classes, and skill stamps appearing on new blocks.
Connection — is the mentee now writing stamps about others? Are they earning their own Beans? This is the cleanest signal because it is binary-detectable on-chain.

The composite effect is that Beane’s framework, which previously had to be measured by post-hoc surveys with all their noise and recall bias, becomes a population-scale longitudinal measurement on the protocol itself.

The Selection Effect

Mentors are economically forced to act like venture investors for skill. Their return depends on whether the mentee actually grows, so they invest selectively, deeply, and over time. A mentor who runs ten shallow checkbox sessions captures less value than one who develops three mentees over years.

The Disadvantage Multiplier

Without correction, deferred-dividend Beans entrench inequality — mentors avoid hard-to-bet-on workers. The Bean Chain corrects this with a disadvantage multiplier: payouts are weighted inversely by the mentee’s baseline. A Bean for mentoring a worker with thin chain history who then trajectories upward pays out more than a Bean for mentoring a worker with strong starter signal. The protocol explicitly rewards mentors who took chances on hard-to-bet-on workers. Structurally analogous to social-impact venture capital.

Regulatory Framing

Bean Chains are structurally identical to deferred-compensation-with-clawback frameworks already operated by every major bank under post-2008 regulatory regimes (Dodd-Frank, EU CRD, APRA prudential standards). Outcome-based incentive compensation, malus provisions, and ex-post risk adjustment are well-understood legal categories. Bean Chain dividends are functionally equivalent to a multi-year performance bonus subject to clawback for inadequate performance — with the difference that the “performance” being measured is the protégé’s trajectory rather than the firm’s revenue. A Tier-1 institution can implement it tomorrow.

6.3.6 Bean Chains as Skill Superannuation (the Institutional Frame)

The deferred-dividend Bean Chain has a financial-instrument analogue that makes it instantly legible to institutional adopters: it is structurally equivalent to superannuation, but for skill transfer rather than for retirement income. This is the framing that turns the protocol from a clever idea into a category that finance ministries, HR departments, and union leadership already understand and have legal scaffolding for.

Bean Chains map onto super:

Every withdrawal costs Beans ≅ super-guarantee percentage
Beans deposit into a Bean Chain ≅ super fund
The Bean Chain “invests” the stake across mentorship dyads ≅ fund’s underlying assets are mentee skill-vectors
Compounding produces a long-dated dividend stream ≅ skill-transfer pension paid in conversion-rate bonus and reputation as a high-quality mentor

The structural claim: HOP makes skill transfer a financial asset class. Currently, skill transfer is unmeasured, unrewarded, undifferentiated. Under HOP, it becomes a portfolio. A senior worker’s Bean Chain holdings are functionally equivalent to a super balance.

Mature retirement-saving systems are typically subsidised through tax concessions in the tens of billions annually, justified by retirement saving as a public good. The same argument applies, with even more force, to long-term skill investment. Bean Chain dividends should attract similar instruments. This is the v0.3 lobbying position; the framing makes HOP economically irresistible to Treasury once they understand it.

6.4 Recursive Decomposition and the Tax Cascade

Work is fractal. A $10,000 bead posted at the root level can be claimed by a worker who immediately reposts decomposed sub-beads — nine $1,000 children, each of which can be further decomposed into $100 grandchildren. Each block holds a parent reference, making the entire decomposition tree traversable.

Tax accumulates at every level of the tree. The $10,000 off-ramps to the winning worker minus their chain’s tax. They post $9,000 worth of bids, each of which off-ramps to its winner minus that chain’s tax. By the time tax reaches the leaf nodes, three off-ramp events may have occurred — one at each tree level.

This sounds like a lot of tax until compared to the existing platform stack. A startup hiring contractors through Upwork pays Upwork 10–20%; the contractor pays Upwork another fee on payouts; the contractor’s bank charges international transfer fees; informal sub-subcontracting adds further unattested friction layers. HOP makes the tax visible at every layer, and crucially each layer’s tax funds that layer’s commons.

Each level also contributes to its own worker’s Skillchain. The leaf-node implementer earns a character block reflecting their specific contribution. The mid-tier integrator earns a character block reflecting integration and project management work. The root worker earns a character block reflecting whole-project shaping and quality assurance. One $10,000 contract produces three different kinds of skill demonstration on three different Skillchains. The decomposition is positive-sum at every level.

6.5 The Dignity Floor (Basal Rate)

The Basal Rate Law (§1) requires a dignity floor below which participants cannot fall. The protocol implements the floor in three places:

Bootstrap grants. A new participant joining a Workchain receives a small starter balance — enough to make a few cold bids and prove themselves.
Service base costs. Every service has an algorithmically protected base cost covering time, physical energy, and materials. Surge pricing below this floor is structurally prevented.
Float, not cliff. Reputation degrades smoothly, not categorically. A worker with low recent reputation can still receive messages, still respond, still do work to rebuild — they simply cannot initiate aggressive cold outreach. They dim; they do not disappear.

The anti-pattern this design refuses is Madogiwazoku — “the tribe that sits by the windows” — the Japanese practice of keeping employees on payroll with status but no work, often progressing into oidashibeya (windowless expulsion rooms). HOP’s float architecture has no integer state to fail into. A worker whose role is automated transitions to mentorship, advisory work, partial capacity, or pivots to growth blocks. They never hit zero.

6.6 Cross-Chain Federation

A rideshare driver in one country who wants to spend their chain-tokens with a manufacturer in another country has two paths:

Naive path: off-ramp chain-tokens to local fiat, send fiat across borders, on-ramp to the destination chain’s tokens. Two off-ramp taxes plus international transfer friction. Expensive.
Federated path: direct chain-to-chain swap if both chains have signed a federation treaty. A single negotiated tax, shared between chains. Cheaper, but requires explicit treaty signature.

Federation treaties become a first-class object of geopolitics. A nation that writes the standard treaty template — the Zollverein move — captures the network position without owning the network.

6.7 The Forking Rule (Anti-Monopoly by Construction)

This subsection defines the protocol’s most important anti-capture mechanism. Without it, the first four Laws are aspirational. With it, they are enforced by the structural possibility of exit. The Forking Rule is the Law that makes the other Laws work.

The rule is simple to state: any open chain can be forked. A fork is a copy of the chain’s state plus a substitute set of services, run by someone else, with a different governance configuration. Workers can move to the fork unconditionally, taking their Skillchains with them. If the fork’s services are good enough, workers migrate. If they are not, workers stay. The market clears.

This means the off-ramp tax rate, the validator quality, the dispute-resolution policy, the Bean issuance liberality, and every other piece of chain governance are continuously priced by the threat of exit. Uber charges 30% because it owns the network effect — drivers cannot leave without losing their passenger pool, their reputation, their work history, their dispute-resolution access. In HOP, the passenger pool is on the public chain. The reputation lives on the worker’s Skillchain. The work history is portable by construction. The switching cost is zero by design. A chain that charges 30% must either deliver 30% worth of services or watch a 12% fork take its workers.

Three things follow:

Platforms are not killed; they are priced honestly. If a chain’s matching algorithms are genuinely excellent, its insurance infrastructure works, its dispute resolution is fast — then 30% might be too high but 15% might be exactly right. Someone forks at 10%, discovers they cannot run the matching well enough at that margin, and workers come back. The fork failed, which proves the original chain was providing real value.
Forking is the natural complement to portability. Skillchain portability without forking allows oligopoly collusion on a high-tax equilibrium. Forking destroys that equilibrium because any single defector can launch a low-tax competitor.
Forking enforces all four substantive Laws. A chain that violates Neutrality is a chain workers will fork away from. A chain that fails the Learning law is one whose workers’ growth-block trajectories will outpace its design. A chain that does not provide a Basal Rate floor will be forked. A chain that violates Privacy will be forked.

The literature on platform competition has been arguing for exactly this set of properties as antitrust correction. Easterbrook (1984) argued that anti-competitive practices and monopoly rents create incentives for market entry; the Forking Rule operationalises that argument structurally. Entry is always one fork away.

The fork retains the Skillchain. This is the load-bearing technical detail. When a worker forks (or migrates to a fork), their Skillchain is unchanged — they keep every character block, every Bean, every stamp, every growth declaration. The fork inherits the worker’s reputation; the original chain loses access to it. Compare to LinkedIn, where leaving means rebuilding your network from scratch; in HOP, leaving means changing the validator URL in your Skill-Agent configuration.

Mentorship Specifically — Voluntary at Chain Level

Mentorship participation is voluntary at the chain level. Each Workchain’s Anchor Block declares whether Bean issuance is opt-in or required. A Tier-1 institution’s chain may declare senior staff must contribute Beans on every withdrawal (contractually mandatory, like super, anchored in the employment contract). A rideshare cooperative may declare it optional. Workers choose which chains to work for partly based on Bean policy.

The Forking Rule keeps this voluntary at the meta-level: a chain whose Bean policy workers find too coercive will see workers fork to a chain with a different policy. A chain whose Bean policy workers find inadequate will see them fork to a chain with stronger participation requirements. The market clears on Bean policy the same way it clears on tax rate.

6.8 What Running a Chain Actually Costs

The Forking Rule prices chain governance honestly, but only if the actual costs of running a chain are visible:

Validator inference. Every character block that requires attestation incurs validator LLM inference cost.
Worker-Agent inference for evaluation. When a posting attracts ten bids of six blocks each, the Worker-Agent must reason over sixty character blocks.
Dispute resolution. When a worker and poster disagree about whether work was completed correctly, someone with authority must adjudicate.
Namespace anchoring. Workchains periodically anchor Merkle roots to a public Data Availability layer.
Bean issuance against future revenue. Beans cost the chain real revenue at off-ramp time.
Bootstrap grants. New participants receive starter balances per the Basal Rate Law.
Agent hosting. Workers without their own hardware run their Skill-Agents as hosted services.
Governance overhead. Chain operators, dispute panels, validator-class managers, namespace key custody, security audits, incident response.
Cryptographic infrastructure. Key management, signature verification at scale, BBS+ proof generation and verification.
Sybil-resistance and trust-matrix maintenance. The Christiano-style trust matrix and Bean Chain longitudinal measurement both require ongoing computation.

A well-run chain at moderate scale (100,000 active workers, 10,000 postings per day, modest dispute load) probably operates at a real cost in the range of 8–15% of throughput. A chain charging 30% is either operating at much larger scale, providing premium services, or extracting rent — in which case forking will arrive shortly.

6.9 Government as Bean Contributor

The protocol’s relationship with government is asymmetric in a way that matters: government can deposit value into chains without being able to capture them.

The current model: government identifies a skills shortage. Government allocates funds. Government tenders a contract to a training provider. Provider captures rents. Provider delivers a curriculum that may or may not produce the skills the economy needs. Outcomes are measured by completion certificates, weakly correlated with actual capability. The fiscal cost is tens of billions annually in any single mature economy, and the return is famously hard to evaluate.

Under HOP, the government has a much better instrument:

Government identifies priority skills (AI literacy in regional or under-served areas, infrastructure-trade transitions, aged-care competency, sovereign-AI engineering).
Government deposits Beans into eligible Bean Chains with eligibility predicates: “this Bean credit is available to mentors whose mentees’ skill-vector trajectories move into [priority direction].”
Mentors who deliver eligible mentorship receive the government-contributed Beans on top of any chain-issued Beans.
Settlement runs the same vector measurement — if the mentee’s trajectory does not actually move in the priority direction, the government Bean is clawed back.

The government is not running the chain. The government is not setting tax rates. The government is depositing into a measurement-validated outcome instrument. Structurally identical to outcome-based government bonds (Social Impact Bonds), but with the outcome being skill-vector trajectory rather than reduced reoffending or improved health metrics.

The Forking Rule keeps this safe. If a government’s eligibility criteria are bad, chains can fork to versions that don’t accept that government’s Beans. If a government tries to capture chain governance through the deposit mechanism, the chain’s workers can migrate.

Three further consequences:

The training-subsidy budget becomes auditable. Currently, $30 billion in Australian skills funding produces poorly-measured outcomes. Under HOP, the government can publish exactly which Beans were issued, which mentees’ trajectories moved in the priority direction, which mentors received settlement, and how much was clawed back. The fiscal accountability is dramatic.
Multiple governments can co-deposit into the same chain. Federal and state. National and supranational. ASEAN and AU and EU all depositing into a federated rural-skills Bean Chain, each with its own eligibility predicates, all settling against the same measurement infrastructure.
The “training Bean” is the right instrument for population-scale upskilling responses to AI displacement. When a sector’s workers face displacement, government deposits priority-direction Beans into the relevant chains, mentors hunt for displaced workers as mentees (because the disadvantage multiplier rewards it), and the displaced workers’ Skill-Agents surface the matched mentors. Retraining happens via the protocol, with measured outcomes, at the natural geometry of the labour market.

This is also the cleanest answer to the question that “let’s just pay the bees for honey” was reaching at: the AI economy will produce surplus value at population scale, but most of it will be evaporative unless an instrument exists that crystallises it as growth in human capability. Beans are that instrument; Bean Chains are the settlement layer; government Bean deposits are the public-finance interface to it.