When are prediction markets most helpful? Evercore ISI has a formula

Evercore ISI says prediction markets are most reliable when contracts are high-volume, near expiration, and tied to clear, objective outcomes; only about 8% of Kalshi and Polymarket events exceed $1 million in volume, while nearly 60% of live markets have under $1,000 traded. The firm argues prediction markets can be useful in chaotic macro, election, and geopolitical events, but thin liquidity, ambiguity, and trader motives can distort prices. It also notes trading volume surged in fall 2025 after institutional attention and the 2024 CFTC approval of election-related contracts on Kalshi.

Analysis

The market is implicitly being split into two businesses: liquid, near-dated event contracts that behave more like a tradable probability surface, and long-dated or thin contracts that behave more like sentiment boards. That distinction matters because the utility is not in “forecasting” per se, but in producing a high-frequency read on positioning and fear; the best alpha may come from using those prices as an input into event-risk hedges, not as a standalone oracle. The thinness of most contracts also creates a structural microstructure edge: large informed capital can move prices enough to create false signal, which means the most crowded headlines may actually be the least trustworthy despite the apparent liquidity.

Second-order, the biggest beneficiaries are not just the prediction venues but any data/market infrastructure stack that can package event probabilities into tradable or benchmarkable products. If prediction markets become a recognized sentiment primitive, the spillover value likely accrues to brokers, API distributors, and media platforms that can drive participation and monetize engagement, while traditional polling and survey providers face further share erosion in fast-moving geopolitical or election windows. The more ambiguous the contract language, the more the market becomes a dispute over semantics rather than outcomes, which caps its usefulness for institutional decision-making and should restrain long-duration adoption by serious hedgers.

The key catalyst is regulatory normalization: if additional event categories get approved or settlement standards tighten, volume can compound quickly because liquidity begets liquidity. The counter-risk is reputational and legal: one high-profile failure to resolve cleanly, or evidence that a small number of traders are repeatedly distorting thin books, could sharply reduce institutional trust over the next 3-6 months. Consensus is likely overestimating how much of the current growth is durable demand versus headline-driven novelty; the real adoption curve will depend on whether these markets can be embedded into risk workflows instead of remaining a side-channel for retail attention.

Key Decisions for Investors

• Avoid chasing the prediction-market headline names on momentum; instead wait for a post-hype pullback in platform-linked beneficiaries and only add on confirmed volume growth over the next 1-2 quarters, since thin liquidity can reverse quickly if one contract blows up.
• Long infrastructure/fintech rails that benefit from event-volume growth, funded by a short basket of traditional polling/survey exposure if liquid; the trade works best if regulatory approvals broaden contract breadth over the next 6-12 months.
• Buy short-dated event-volatility hedges around major political or geopolitical catalysts using options on broad market hedges (e.g., SPY puts into binary event windows) because these markets are likely to sharpen near-term consensus and amplify repricing speed.
• If looking for a cleaner relative-value expression, pair long data-distribution/market-analytics beneficiaries vs. short legacy media names exposed to weaker election-cycle monetization; the adoption thesis is more about traffic and workflow integration than outright forecast accuracy.
• Do not hold any directional position based solely on low-volume contract moves; require at least $100k+ notional depth and 1-2 weeks of stable participation before treating a prediction-market signal as investable.