9.0 KiB
Flow Bonds
A Dynamic Outcome Incentivization Primitive
Flow bonds replace binary prediction market bets with continuous, streaming settlement tied to measurable real-world outcomes. Instead of betting YES/NO on whether something crosses a threshold, participants commit capital expressing beliefs about continuous variables and earn (or lose) yield every period proportional to how right they are.
When the outcome being tracked is a public good, the people betting and the people doing the work can be the same people — turning "market manipulation" into aligned incentive to create positive change.
Core Mechanism
How It Works
- An outcome sponsor defines a measurable variable and commits funding to a flow bond pool
- Participants commit capital expressing directional beliefs (long = outcome improves, short = it won't)
- An oracle reports the outcome metric at regular intervals
- Each period, money flows between participants proportional to how the metric moved — continuously, not at a single resolution event
- Participants can exit anytime, withdrawing remaining collateral +/- accumulated flows
The Three Money Flows
| Flow | Source | Description |
|---|---|---|
| Zero-sum | Participants | Losers pay winners each period, proportional to outcome movement. The perp-futures-like base layer. |
| Sponsor subsidy | Outcome sponsors | Streamed into the pool, distributed to accurate forecasters. Makes the market positive-sum. |
| Agency premium | Participant effort | Those who can influence the outcome have an informational edge. Unlike insider trading, this "manipulation" is the entire point. |
Flow Bonds vs. Perpetual Futures
| Dimension | Perp Futures | Flow Bonds |
|---|---|---|
| Underlying | Asset prices only | Any measurable variable (air quality, tree cover, GDP, soil carbon...) |
| Settlement | Funding rate every 8h, anchors to spot price | Continuous flow per oracle update, anchors to measured reality |
| Purpose | Speculation & hedging | Forecast accuracy + outcome incentivization |
| Funding source | Zero-sum: longs pay shorts or vice versa | Zero-sum base + positive-sum subsidy from outcome sponsors |
| Leverage | High (10-100x common) | Low/none — collateral is commitment, not margin for leverage |
| Participant influence | Manipulation (bad) | Agency (good — doing the work IS the edge) |
| Oracle | Price feeds (Chainlink, Pyth) | Outcome oracles (sensors, satellites, public data, attestations) |
| Credit risk | Margin + liquidation engine | Collateral buffer + gradual liquidation (no sudden blowups) |
Key similarity: Both use continuous settlement to avoid expiry/rollover. Both transfer money between longs and shorts each period.
Key difference: Perps exist to track a price. Flow bonds exist to incentivize an outcome. The subsidy layer from outcome sponsors makes flow bonds positive-sum, meaning accurate forecasters earn yield even in a flat market.
Worked Example: Urban Reforestation
Setup
The City of Vienna commits 200,000 USDC over 2 years, streaming into a flow bond pool tied to "urban tree canopy coverage %" as measured by quarterly satellite imagery (Copernicus/Sentinel, publicly verifiable).
Positions
- Green Guild (a local tree-planting cooperative) commits 50,000 USDC long. They believe canopy will increase because they intend to plant 5,000 trees.
- A hedge fund commits 30,000 USDC short. They're skeptical — the city has failed on green promises before.
- Individual residents commit smaller amounts long, signaling community support.
Settlement
- Q1: Canopy +0.4%. Shorts pay longs. Green Guild earns ~2,000 from zero-sum pool + ~4,000 from sponsor stream. Hedge fund loses ~1,200.
- Q2: Canopy +0.8%. Larger flow to longs. Green Guild has earned back 15% of commitment.
- Q3: Canopy -0.1% (drought). Flow reverses — longs pay shorts. Green Guild loses a little.
- Q4: Canopy +1.2% (autumn rains). Big flow to longs. Hedge fund's collateral down to 40% — they exit.
Outcome
After 2 years, Green Guild has earned a 40% return on committed capital — funded by shorts who were wrong and by the city's outcome sponsorship. The city got 5,000 new trees and market-verified measurement of their green infrastructure progress. The hedge fund provided valuable price discovery and exited with a managed loss. Residents earned modest yields for backing the initiative.
The Inversion: Why "Manipulation" Is the Point
Traditional prediction markets treat participant influence on outcomes as a bug. But if the outcome is a public good, manipulation is the point.
- You want people to bet on reforestation and then go plant trees
- You want people to bet on clean air and then push for better policy
- You want people to bet on education outcomes and then teach
Flow bonds make this viable because of continuous settlement. A binary bet on "will 10,000 trees be planted by 2027" means you sit and wait. A flow bond on "urban tree canopy coverage" means you earn yield every period that coverage increases. The incentive to act is immediate and ongoing.
What This Fixes
| Existing mechanism | Problem | Flow bonds fix |
|---|---|---|
| Binary prediction markets | Distort continuous phenomena into yes/no; deferred payout kills feedback loop | Continuous variable, continuous settlement |
| Social impact bonds | Institutional, binary, slow, expensive to structure | Democratized, continuous, self-executing |
| Retroactive public goods funding | Requires subjective post-hoc judgment | Prospective, market-priced, continuous reward |
| Carbon credits | Verification theater, one-time certification | Tied to measured outcomes continuously; if the forest burns, the flow reverses |
Integration with rStack
Flow bonds connect naturally to several rStack primitives:
rNetwork — Trust & Delegation
Delegated trust as collateral weighting. A highly-trusted community member's position carries more signal (and potentially more sponsor-funded yield) than an anonymous speculator. Trust scores become reputation collateral.
rcart — Payment Infrastructure
Existing payment request/QR system and wallet integration provide collateral commitment and payout rails. Flow bond positions could be created and settled through the same crypto payment flows.
rVote / rChoices — Governance
A DAO uses rVote to decide which outcomes to sponsor, committing treasury funds to flow bond pools. rChoices lets communities rank which public goods metrics matter most.
EncryptID — Identity & Attestation
DID-based identity enables reputation-weighted positions and prevents sybil attacks. Attestation flows serve as oracle inputs — community members attesting to on-the-ground outcomes.
Proposed: rBonds Module
A dedicated module housing:
- Bond creation interface — outcome sponsors define metrics and commit funding
- Position manager — participants commit collateral and track yields
- Oracle registry — connecting to data feeds (satellite, sensor, API, attestation)
- Settlement engine — processes continuous flows each period
- Visualization — rNetwork's 3D graph shows capital flow between participants, sized by position and colored by outcome performance
Open Questions
- Oracle design: What's the trust model for outcome data? Automated feeds (satellites, sensors) are trustworthy but limited. Attestation-based oracles are flexible but gameable. Likely need a hybrid with dispute resolution.
- Liquidity bootstrapping: How do you get the first shorts into a "bet on better futures" market? Shorts provide a critical service (price discovery, keeping optimists honest) and should be framed as such.
- Regulatory framing: Is this a derivative, a donation, an impact bond, or a prediction market? The sponsor subsidy layer might qualify it as outcomes-based contracting.
- Settlement frequency: How often does the oracle update? May vary by metric (air quality hourly, tree canopy quarterly).
- Composability: Can flow bond positions be tokenized and traded? This creates a secondary market for "impact exposure."
- Collateral types: Can non-financial commitments (labor, materials, expertise) count as collateral alongside capital?
Origin
This concept emerged from a discussion about the aesthetic and structural limitations of binary prediction markets — specifically, that shoehorning continuous outcomes into YES/NO contracts (a) distorts the underlying signal, (b) appeals primarily to gambling instincts, and (c) solves the credit problem only by sacrificing expressiveness. Flow bonds attempt to keep the credit-risk benefits of upfront collateral while restoring the continuous, dynamic nature of real-world outcomes — and adding the crucial insight that participant agency over outcomes is a feature, not a bug.
Built in the spirit of P4P — rstack