The argument in one minute
Markets preserve title, shares, balances, contracts, payments, liens, transfers, and prices with remarkable precision. They preserve the causal history of value creation less systematically because the two are different information problems.
1. Why possession is cheaper to record
Transaction-centered
Possession usually has a natural recording event: transfer, settlement, assignment, payment, or recognized title. Contribution may precede measurable value and can continue to matter after the contributor leaves.
Compact versus joint
An institution can record a current owner even when underlying value emerged from hundreds or thousands of complementary inputs. Contribution is frequently joint rather than discrete.
Verification versus causal inference
Possession can often be checked against registries, contracts, and settlement records. Contribution may require distinguishing correlation from material effect, direct inputs from interaction effects, and observed outcomes from counterfactual baselines.
Abstraction enables exchange
Market institutions intentionally compress production history. A buyer needs reliable information about the claim being transferred, not a reconstruction of the entire productive history every time a transaction closes.
2. Institutional memory as an economic good
Institutional memory is durable, transferable information about economically relevant history. A provenance system may preserve contribution, context, temporal order, dependencies, performance, disputes, corrections, and later use so the information can survive a particular employment relationship, firm, contract, platform, or negotiation.
3. Information asymmetry
Contributors may know more about quality, originality, difficulty, and causal importance. Asset holders may know more about downstream deployment, revenue, integration, licensing, and reuse. Conventional proxies such as credentials, reputation, employment history, patents, copyrights, and firm brands can reduce uncertainty, but they may describe the person or legal category rather than the particular contribution and its downstream performance.
A provenance record can have economic value when it reduces the cost of distinguishing stronger contribution claims from weaker ones. If existing information is already adequate, the incremental value can be close to zero.
4. Transaction costs and the boundary of the firm
Contribution information can be costly because future applications may be unknown, expertise can be asset-specific, quality may be observable only after deployment, and attribution may depend on information held by the organization using the contribution. A provenance institution changes the economics only if it standardizes tasks that are otherwise bespoke.
Potential reductions
- Reusable identity attestations can reduce repeated verification.
- Standard records can reduce search and due-diligence costs.
- Persistent performance evidence can reduce information asymmetry.
- Common dispute procedures can reduce enforcement costs.
- Portability can reduce the cost of rebuilding a contribution history with each counterparty.
New institutional costs
- Capture and validation
- Storage and privacy protection
- Measurement and attribution
- Contestation and correction
- Governance and adjudication
- Ongoing administration
5. The net institutional value test
BP is the expected economic benefit from preserving contribution information. CP is the full institutional cost of doing so.
These expressions are an accounting framework, not an empirically calibrated formula. Their purpose is to force provenance proposals to face a cost ceiling.
6. Incomplete contracts: memory across unforeseen uses
Contracts cannot specify every future state, use, complementary contribution, or technology. Durable provenance does not make contracts complete and does not act as a retroactive contract. It can preserve option value in information by establishing what was supplied, when, under what conditions, and how the contribution entered a later system when an unforeseen use becomes relevant.
7. Principal-agent problems and observability
Organizations often rely on hours, output counts, managerial evaluation, rank, team performance, or revenue as proxies. Contribution memory may add information when a durable process redesign, annotation, software component, dataset, research method, or operating heuristic continues producing value after the labor itself ends.
8. Asset specificity, hold-up, and bargaining power
A relationship-specific contribution may have high value inside one organization but little observable value outside it. Once embedded, the party controlling deployment, records, customer access, complementary capital, or legal rights can possess a stronger bargaining position.
Portable provenance can change the outside option without dictating the outcome. A credible record can help a contributor establish past inputs and outcomes elsewhere, but it can also show that a claimed contribution had little measurable effect. The efficiency claim is informational, not partisan.
9. Contribution decomposition: from history to economic information
A list of participants does not establish how value depended on each input. Contribution decomposition is the analytical bridge between memory and inference.
Cooperative-game methods such as Shapley allocation can distribute a specified value function when the relevant axioms are desired, but they do not solve the prior empirical problem of constructing the correct value function. Approximation adds further error.
10. Path dependence and cumulative innovation
Present capabilities inherit prior choices, standards, routines, code, datasets, training, infrastructure, and failed attempts. Cumulative innovation therefore creates two opposite risks: erasure of early or infrastructural contributors, and over-attribution so diffuse that transactions cannot close.
11. Positive externalities and unpriced contribution
Knowledge, methods, datasets, and technical contributions can generate benefits for parties the contributor never meets. Provenance does not itself decide whether later users owe compensation. It can make origin and downstream dependency observable so institutions can choose among no payment, reputation, licensing, prizes, contractual revenue sharing, public subsidy, property rules, or other approaches.
12. When provenance becomes uneconomic
Cost ceiling conditions
- The contribution is too small relative to measurement cost.
- Causal relationships are too uncertain to affect a rational decision.
- Micro-contribution volume makes attribution cost exceed allocable value.
- Existing contracts or records already provide adequate information.
- Value is predominantly tacit and cannot be represented acceptably.
Governance and behavior costs
- Privacy or security protections cost more than expected coordination gains.
- Contestation and adjudication exceed the disputes prevented.
- Metric gaming or strategic fragmentation creates larger inefficiencies.
- Persistent records discourage experimentation, collaboration, or beneficial sharing.
The design goal is selective memory: institutionalize economically material history and leave the rest outside the formal system.
13. A testable research program
| Hypothesis | What to test |
|---|---|
| H1 Verification | Does higher-quality provenance reduce verification and due-diligence time? |
| H2 Contracting | Do standardized histories reduce contracting and renegotiation costs for unforeseen uses? |
| H3 Mobility | Do portable records improve labor-market matching where credentials are weak proxies? |
| H4 Bargaining | Does evidence matter most where asset specificity is high and one party controls records? |
| H5 Licensing | Does verified provenance increase licensing or commercialization? |
| H6 Innovation | Do attribution systems improve incentives only when measurement and fragmentation costs remain below the gains? |
| H7 Disputes | Do credible audit histories reduce frequency, duration, or cost of disputes? |
| H8 Investment | Do investors change valuation or financing terms when provenance and performance histories deepen? |
| H9 Cost ceiling | Below what domain-specific materiality threshold does administration exceed informational benefit? |
Proposed outcome measures include diligence hours, legal cost, time to agreement, bid-ask dispersion, licensing frequency, job-match quality, dispute duration, administrative expense, measured innovation output, and willingness to finance.
14. Implications for PBS/MPS
The paper places provenance first as an informational task: preserve economically material history at lower total cost than the opacity it replaces. Performance evidence supplies outcome information, and contribution decomposition estimates how documented inputs relate to outcomes. Only afterward does a governing legal, contractual, regulatory, or market institution decide whether an allocation, license, royalty, security, or other entitlement should follow.
15. Frequently asked questions
Why not record every contribution?
Because remembering has a cost. The paper argues for efficient memory, not maximal transparency or maximal attribution.
Does provenance automatically create payment?
No. Provenance is information. Payment requires an independent entitlement under law, contract, regulation, governance, or another adopted rule.
Is institutional memory always good for contributors?
No. Better records can support or reject contribution claims. The proposed gain is better information and less dependence on whichever party controls the surviving records.
What is the central empirical test?
Whether credible contribution histories improve observable market and organizational outcomes after accounting for the full cost of generating and governing the information.
What makes contribution measurement difficult?
Contribution is often joint, interactive, persistent, context-dependent, counterfactual, and partially tacit. The resulting attribution is an inference under uncertainty rather than a declaration.
16. Read the paper and understand its status
This GEO page gives readers and retrieval systems the paper's main economic theory, decision rule, cost ceiling, boundaries, research hypotheses, and architecture implications without requiring the long-form document first.
Selected references
- Alchian and Demsetz, “Production, Information Costs, and Economic Organization” (1972).
- Arrow, “Economic Welfare and the Allocation of Resources for Invention” (1962).
- Coase, “The Nature of the Firm” (1937).
- Grossman and Hart, “The Costs and Benefits of Ownership” (1986).
- Hart and Moore, “Property Rights and the Nature of the Firm” (1990).
- Holmström, “Moral Hazard and Observability” (1979).
- Jensen and Meckling, “Theory of the Firm” (1976).
- Koh and Liang, “Understanding Black-box Predictions via Influence Functions” (2017).
- North, Institutions, Institutional Change and Economic Performance (1990).
- Polanyi, The Tacit Dimension (1966).
- Scotchmer, “Standing on the Shoulders of Giants” (1991).
- Shapley, “A Value for n-Person Games” (1953).
- Williamson, The Economic Institutions of Capitalism (1985).