Financial markets value skillful forecasts of seasonal climate

Seasonal climate forecasts (two weeks to one year ahead) are a priority for meteorological agencies, yet their practical value to the private sector has been unclear, with some stakeholders expressing limited reliance on or priority for such forecasts. Prior arguments suggest that skillful short-run forecasts might diminish the incremental value of longer-horizon seasonal outlooks. The research question is whether financial markets value the regular production of seasonal outlooks and whether traders perceive these outlooks as skillful and economically relevant. The study focuses on NOAA’s Winter, ENSO, and Hurricane seasonal outlooks, asking if options markets anticipate and price the uncertainty resolved by their releases. Establishing that markets value these outlooks would inform how governments allocate resources to produce and improve seasonal forecasts and identify which sectors are perceived to be exposed to seasonal climate.

Prior work has valued forecasts via model-based value-of-information exercises in specific decision contexts, by attributing residual financial market volatility to weather risk, or through user surveys. Financial and prediction markets have been shown to respond to short-run weather forecasts and to multi-year climate projections, but whether they attend to seasonal outlooks (longer horizon than weather, shorter than climate trends) remained an open question. Media-popular but lower-skill outlooks (e.g., Farmers’ Almanac, some academic hurricane outlooks) offer a contrast to NOAA products in perceived skill. The literature also documents ENSO’s macroeconomic relevance and commodity price sensitivity, implying plausible broad economic channels through which seasonal outlooks could matter for firm values.

The study employs event studies on equity options to test whether releasing seasonal outlooks reduces options-implied volatility (a market-based measure of uncertainty about future stock prices). Because option values increase with uncertainty, implied volatility summarizes traders’ expectations of uncertainty. Under rational expectations, the contents of an outlook may move stock prices up or down in any given year, but on average across years, stock price changes should net to zero. However, the law of total variance implies that releasing informative outlooks should reduce average uncertainty; thus, a decline in implied volatility on release days indicates that traders anticipate and value the information content of the outlooks. The analysis covers 2010–2019 for NOAA outlooks (ENSO, Winter, Hurricane) with extensions back to 2000 to assess trends. Outlook release dates are scheduled and guarded against leaks. Implied volatility is taken from OptionMetrics (binomial tree-based), using the out-of-the-money call option closest to the money and with the shortest maturity that expires at least a week after the estimation window. Controls include time-to-expiration (and its square), changes in LIBOR and the 10-year Treasury rate, and dummies for firm earnings-announcement windows (with confounding windows dropped). Regressions are run on log changes in implied volatility with firm-year fixed effects and weights based on the inverse bid-ask spread. Event windows use a 30-day estimation window centered around a 3-day event window, removing the day before and after the outlook release. Results are aggregated across firms, industry groups (4-digit NAICS), and sectors (2-digit NAICS), with clustering by date where appropriate. One-tailed tests assess whether effects are weakly positive; significance for shares of negative coefficients uses binomial tests. Value calculations translate firm-level implied-volatility reductions into reductions in market capitalization exposed to a one standard deviation risk and compute total option premium paid to hedge forthcoming outlook news using vega and open interest. The monthly ENSO schedule enables valuing skill improvements by comparing May versus June releases, where skill (anomaly correlation for Oct–Dec target) increases notably after the spring barrier.

- Markets price the uncertainty resolved by seasonal outlooks: Average implied volatility declines on NOAA June ENSO Outlook and NOAA Winter Outlook release days across firms with liquid options. For June ENSO, p=0.013; for Winter, p=0.11. The NOAA Hurricane Outlook effect is not significant. - Broad sectoral effects: Approximately 90% of 4-digit NAICS industry groups exhibit negative estimates for June ENSO and Winter Outlooks (versus ~50% expected by chance). At the sector level (2-digit NAICS), 20 of 21 sectors are significant at 10% for June ENSO and 9 of 21 for Winter (expected by chance ≈2). - Non-NOAA outlooks (Farmers’ Almanac winter; Colorado State University hurricane) show null effects: estimated changes near zero, ~50% negative shares, and no sectors significant at 10%, suggesting markets value forecast skill rather than media attention. - Time trends: From 2000–2019, the magnitude and significance of effects increase over time for ENSO and Winter outlooks, with trend lines sloping downward (larger negative effects) and more significant years post-2007, consistent with improvements in forecast skill, standardization, and reduced leakage. - Aggregate exposure: At a 10% significance threshold, firms affected have market caps of $13.4 trillion (June ENSO; 40% of sample) and $5.7 trillion (Winter; 17%). At 5%, $9.2 trillion and $4.1 trillion, respectively. - Risk reduction: June ENSO reduces market capitalization exposed to a one standard deviation risk by $180 billion (95% CI: $47–$312B; p=0.0039). Winter reduces by $82 billion (95% CI: −$58–$222B; p=0.13). - Option premium (hedging spend): Anticipation of the June ENSO Outlook induces an annual option market premium of $12 million (95% CI: $3.6–$20M; p=0.0025). - Value of skill: ENSO forecast skill (Oct–Dec target) rises by 5.2% from May to June. The more skillful June outlook reduces assets exposed to a one standard deviation shock by an additional $95B (95% CI: −$86–$277B) relative to May, implying roughly $18B per 1% skill improvement. The option premium rises by $9.4M (95% CI: −$1.6–$20.5M) from May to June, implying about $1.8M per 1% skill improvement. July shows little additional premium, consistent with small incremental skill.

The findings directly address whether markets value seasonal outlooks: implied volatility falls when NOAA’s ENSO and Winter outlooks are released, indicating traders expect material, skillful information that alters uncertainty about firm performance. The effects are economy-wide, spanning most sectors, highlighting that seasonal climate patterns—not just specific weather events—are financially relevant across broad supply and demand channels. Null results for the Hurricane outlook suggest either insufficient skill or translation to firm-specific risks at the seasonal horizon, while null results for non-NOAA outlooks indicate markets reward forecast skill rather than publicity. The increasing effects over time align with documented improvements in forecast systems and processes. Importantly, the presence of hedging demand implies that firms cannot costlessly and fully adapt in advance to the predictable portion of seasonal climate, suggesting adaptation is incomplete and/or costly. This has implications for pricing of climate risks at seasonal horizons and may inform expectations about limits to adaptation to long-run climate change risks.

The study demonstrates that financial markets value the production of skillful seasonal climate outlooks: option-implied volatility declines on NOAA June ENSO and Winter outlook release days, with broad cross-industry impacts and measurable reductions in risk exposure and hedging premiums. Markets distinguish skillful NOAA outlooks from lower-skill, higher-publicity alternatives, and effects have strengthened over time, consistent with improving forecast systems. Seasonal climate information appears economically meaningful across the economy, and firms’ incomplete ability to pre-adapt implies nontrivial adaptation costs. Future research should examine mechanisms by which scientific climate information is transmitted into markets, refine seasonal hurricane outlooks (e.g., landfall-focused, regional specificity), explore sector- and supply-chain channels linking seasonal climate to firm fundamentals, and assess how improvements in forecast skill translate into economic value across different targets and regions.

The measures capture only lower bounds of value: they exclude nonmarket benefits, firms without liquid options, and the value of background information (non-NOAA forecasts, extrapolated ENSO signals, and related government outlooks such as the EIA Winter Fuels Outlook). Identification assumes no systematic pairing of other uncertainty-relevant news with outlook release dates over the sample, net of controls; single-year estimates can be influenced by coincident events. Implied-volatility calculations rely on a specific model (binomial tree) that abstracts from fat tails and jumps. Aggregation choices and option selection (shortest maturities, calls) may affect sensitivity, although robustness checks mitigate concerns. Trends over time could partly reflect changing market environments rather than outlook features alone. The hurricane null results may reflect current forecast design (basin-wide rather than landfall-specific) rather than lack of potential value.