Concept · business-and-platforms
Cross-subsidization
Quick definition. Using revenue from commercially profitable PMs to fund liquidity incentives on socially valuable but not-yet-self-sustaining markets · mirroring how newspaper ads subsidized investigative journalism.
Key insights
- Alan Wu's central frame: PM prices are public goods. The benefits (informational signal, hedging tool, accountability) are non-excludable, but the liquidity costs fall on a narrow trader base. Without cross-subsidization, only commercially profitable markets survive · and the most informationally valuable markets (long-tail policy, scientific outcomes, accountability) are often the least profitable.
- The newspaper analogy: classifieds and display ads funded the investigative-journalism beat that society needed but no individual reader would have paid for directly. PMs can do the same · sports and crypto profits subsidize political, scientific, and policy markets.
- Wu's secondary argument: accuracy isn't the only axis of value. Markets serve risk transfer (hedging hurricane or policy exposure) and information accountability functions even when prices drift from pure probability. The cross-subsidy can be justified by these auxiliary functions even where the market isn't perfectly calibrated.
- This connects directly to Hall & Paschal's "Building the Truth Machine" prescription: cross-subsidize political liquidity from sports profits. They find only 1.3% of political markets are liquid enough to be manipulation-resistant · fixing this requires money to flow from where it's earned (sports) to where it's needed (politics).
- The cross-subsidization concept's related concepts (per the on-prediction page) include adverse selection · informed traders systematically extract from MMs in low-volume political markets, and the cross-subsidy is precisely the mechanism for compensating MMs for taking that loss.
- Hedging-as-justification: Wu argues some socially valuable markets (climate, geopolitical risk) exist primarily as risk-transfer mechanisms, not pure forecasting. Cross-subsidies are justified because the social value of available hedges exceeds the narrow trading profits the market generates.
- Implicit critique of pure profit-maximization: platforms that only run sports/crypto markets are not contributing to the public-good case for PMs; they're free-riding on the legitimacy that information-valuable markets provide.
- The mechanism could take multiple forms: explicit subsidies (MM rebates funded from sports rake), platform fee redistribution, or governance-level allocation. Andy Hall's a16z piece implicitly assumes the platform itself does the allocation.
- Cross-subsidization is essentially "Tobin tax on entertainment markets to fund information markets" · a redistributive design choice that platforms can make voluntarily or that regulators could impose.
- Without cross-subsidies, the ghost-market problem (functionSPACE) gets worse over time: as more questions are atomized into thin binary contracts, more markets fall below minimum viable liquidity, and the informational-value half of the PM thesis hollows out.
- Cross-subsidies don't address the manipulation problem (Prosperi documented 23% open-interest concentration); they only address the funding problem. A subsidized market that's still thin can still be manipulated.
In their words
Prediction market prices are public goods whose benefits are non-excludable but whose liquidity costs fall on a narrow trader base.· paraphrased from Alan Wu, *How Prediction Markets Can Ascend*
Profitable markets fund socially valuable ones that can't sustain themselves, the same way newspaper ads funded investigative journalism.· paraphrased from Alan Wu
Where it matters
Cross-subsidization is the missing economic mechanism for making PMs work as information infrastructure rather than pure entertainment. Without it, the platform incentive is to keep adding sports/crypto markets indefinitely; with it, the platform's own sports profits underwrite the long-tail political markets that justify the "truth machine" narrative. For Dekant: continuous-curve markets have a structural cross-subsidy property · a single liquidity pool over a continuous outcome covers the whole curve, so liquidity earned on a popular region's trading volume effectively underwrites the tails of the same distribution without an explicit subsidy mechanism.
Connections
- Liquidity provision · the cost that needs subsidizing
- Adverse selection · the cost MMs face in thin markets
- Long-tail markets · the obvious subsidy target
- Hedging · the auxiliary use case justifying socially valuable markets
- Minimum viable liquidity · the threshold subsidies aim to push markets above
Platforms linked to this concept
- Polymarket · implements · Cross-subsidy across markets on Polymarket
Related concepts
Sources
- How Prediction Markets Can Ascend · Alan Wu · Mar 25 2026 ·