Prolepsia forecastYes / no
Will the S&P 500 fall more than 15% from a peak at any point in 2027?
As of 15 November 2026, Prolepsia puts the chance of yes at 32%.
32%chance of yes
How Prolepsia reasons
The report Prolepsia wrote with this forecast.
How forecasts are made · Track record
TL;DR
The forecast is 32.2% probability of Yes and 67.8% probability of No. No qualifying decline is the more likely outcome, but the risk is substantial. Only daily closing declines of strictly more than 15% from an earlier closing high within 2027 count.
Key evidence
- History puts the event within the normal range of market risk. The historical record contains 25 qualifying years out of 76 during 1950–2025, 17 out of 46 during 1980–2025, and 11 out of 26 during 2000–2025. These overlapping periods describe the same market history, not independent evidence (historical daily prices).
- Earnings and valuation counter the bearish case. FactSet’s November 13 report showed 31.9% blended third-quarter earnings growth with 91% of companies reporting. The forward price-to-earnings ratio was 18.3, below its reported five- and ten-year averages. Forecast 2027 earnings growth of 16.2% remains an expectation, not a realized result (FactSet).
- Current conditions show turbulence rather than an established crisis. VIX closed at 19.30 on November 13 (Cboe). High-yield credit spreads narrowed from 3.48 percentage points on November 5 to 3.20 on November 12 (FRED). October unemployment was 4.1%, with payrolls up 165,000 (BLS).
- Inflation and financing remain credible threats. October headline and core CPI rose 0.4% and 0.3% month over month, respectively (BLS). The Federal Reserve also identified large advanced-computing financing commitments that could amplify an earnings disappointment (Financial Stability Report).
Uncertainties
The central unknown is the economic and volatility regime during 2027. Strong current earnings do not prevent later estimate cuts, and contained aggregate credit spreads do not rule out stress in particular borrowers.
Presidential-cycle history offers a modest counterweight: only two of 19 comparable third-cycle years qualified. That sample is too small to justify treating the pattern as a dependable rule. The long historical record also has early-data limitations and lacks a fully independent daily-history validation.
Finally, this is a path-dependent event, not a forecast of the annual return. One qualifying close is sufficient, even if the index subsequently recovers. Conversely, an early-2027 decline measured only against a 2026 peak does not qualify.
Read the full report
TL;DR
The forecast is 32.2% probability of Yes and 67.8% probability of No. No qualifying decline is the more likely outcome, but a decline of strictly more than 15% remains a substantial risk. The decisive measurement is a daily closing loss from an earlier 2027 closing high, not the annual return or the distance below a 2026 record.
Context
As of the November 15, 2026 forecast date, the S&P 500 remained near its record closing high despite recent turbulence. It closed at 7,561.67 on November 13, compared with 7,818.93 on October 6. Neither level carries into the question’s reference peak: the running closing maximum starts again with the first trading close of 2027 (S&P DJI-sourced FRED series).
The event requires one official daily close strictly more than 15% below the highest preceding close within 2027. Intraday lows and dividends do not count. The reference peak need not be an all-time record, and a later recovery does not reverse a Yes outcome. This makes the question about the market’s path during the year, rather than where it finishes.
Evidence
The historical backbone is the frequency of this exact event, measured using daily price-index closes and a fresh running maximum each calendar year. During 1950–2025, 25 of 76 years qualified, a historical frequency of 32.9%. During 1980–2025, 17 of 46 qualified, or 37.0%; during 2000–2025, 11 of 26 qualified, or 42.3%. The shorter recent record looks riskier, but it contains fewer years and several clustered crises. These windows overlap and cannot be counted as separate confirmations (historical daily-price data).
The full decade breakdown shows why a single recent period can mislead. Qualifying years are unevenly distributed, rather than arriving at a stable rate. The following counts use the same within-year closing-price definition (daily-price history).
| Period | Qualifying years | Years observed |
|---|---|---|
| 1950–1959 | 1 | 10 |
| 1960–1969 | 3 | 10 |
| 1970–1979 | 4 | 10 |
| 1980–1989 | 4 | 10 |
| 1990–1999 | 2 | 10 |
| 2000–2009 | 5 | 10 |
| 2010–2019 | 3 | 10 |
| 2020–2025 | 3 | 6 |
The historical threshold also matters. A decline exceeding 10% occurred in 41 of the 76 years, while a decline exceeding 20% occurred in 13. The specified event sits between a relatively common correction and a deeper bear-market decline. It does not require an economic collapse, but it is not a routine feature of every year (historical daily prices).
Presidential-cycle history points in a more reassuring direction. The year 2027 is the third year of the presidential cycle. Only two of the 19 corresponding years in the postwar sample exceeded the threshold. The complete reported third-cycle history follows; the figures are maximum within-year closing drawdowns, not annual returns (historical daily prices).
| Year | Maximum closing drawdown |
|---|---|
| 1951 | 8.1% |
| 1955 | 10.6% |
| 1959 | 9.2% |
| 1963 | 6.5% |
| 1967 | 6.6% |
| 1971 | 13.9% |
| 1975 | 14.1% |
| 1979 | 10.3% |
| 1983 | 6.9% |
| 1987 | 33.5% |
| 1991 | 5.6% |
| 1995 | 2.5% |
| 1999 | 12.1% |
| 2003 | 14.1% |
| 2007 | 10.1% |
| 2011 | 19.4% |
| 2015 | 12.4% |
| 2019 | 6.8% |
| 2023 | 10.3% |
That pattern deserves limited weight. Nineteen observations are a small sample, and the calendar does not establish a causal protection against losses. The two exceptions also show that favorable seasonality offers no shield against a large shock. The useful conclusion is a modest counterweight to the broad historical risk, not a reason to dismiss it.
Current earnings provide a stronger economic counterweight. FactSet’s November 13 report showed blended third-quarter earnings growth of 31.9%, with 91% of companies reporting. The forward 12-month price-to-earnings ratio was 18.3, below the reported five-year average of 19.8 and ten-year average of 19.1. FactSet projected calendar-2027 earnings growth of 16.2%. The reported results support the market; the projection remains exposed to revisions (FactSet Earnings Insight).
The same report adds nuance to apparently weak earnings surprises. The aggregate third-quarter surprise was 3.1%, but excluding three unusually influential company misses raised it to 6.7%; special charges contributed to those misses. This weakens the claim that the headline figure proves a broad operating slowdown (FactSet).
Market stress indicators are mixed. VIX rose to 28.18 on November 5, then fell to 18.70 on November 12; Cboe reported 19.30 on November 13. That sequence shows a volatility burst followed by a retreat, not a persistent crisis reading. VIX measures near-term option-implied volatility. It is not a probability of this annual drawdown event (FRED VIX history; Cboe).
Credit spreads tell a similar story. The high-yield option-adjusted spread rose from 2.78 percentage points on August 3 to 3.48 on November 5, before narrowing to 3.20 on November 12. Financing risk increased over the broader window, but the retreat from the peak argues against extrapolating immediate distress (FRED credit-spread series). October unemployment of 4.1% and payroll growth of 165,000 also argue against describing an employment collapse as already underway (BLS employment release).
Inflation is less reassuring. October headline CPI rose 0.4% month over month and core CPI rose 0.3%, both seasonally adjusted. Those readings create a route to valuation pressure if inflation constrains monetary-policy support. The usual October year-over-year comparisons were unavailable because the corresponding October 2025 observations were missing; an estimated annual inflation figure should not be treated as an official reading (BLS CPI release).
There are also financing vulnerabilities beneath the aggregate data. The Federal Reserve reported more than $680 billion of advanced-computing-related debt issuance through October 2026, compared with $130 billion during 2023. It identified more than $1 trillion in not-yet-started leases and more than $2 trillion in guarantees and other commitments. These undiscounted obligations cannot simply be added to balance-sheet debt. They nevertheless describe channels through which a disappointing investment cycle could spread losses beyond individual technology companies; strong hyperscaler cash flows are a mitigating factor (Financial Stability Report).
The lending evidence supports watching that channel without declaring a general credit seizure. Domestic banks reported tightening across all surveyed categories of lending to nondepository financial institutions, while standards for large and middle-market commercial and industrial borrowers were basically unchanged. I read this as localized caution that could matter before aggregate measures deteriorate (October lending survey).
What's non-obvious
The calendar reset changes the intuitive reading of current market weakness. A sell-off that begins from a 2026 high does not automatically give this event a head start. But the index does not need to regain that old record either: an early-2027 close establishes a new eligible reference level. The relevant question is how far prices subsequently fall from the highest preceding close that year.
A second distinction is between evidence of vulnerability and evidence of an approaching break. Strong earnings and contained aggregate spreads coexist with concentrated financing commitments and tighter lending to nonbanks. Neither side cancels the other. The evidence supports a meaningful drawdown risk without supporting a crisis-dominated forecast. Recent November turbulence also has limited demonstrated power to predict the following calendar year’s maximum drawdown.
Uncertainties
The largest unknown is the economic and volatility regime during 2027. An earnings reversal, renewed inflation pressure, a credit event or an external shock could overwhelm today’s supportive conditions. Continued profit growth and stable financing could instead keep declines below the threshold. Current observations establish the starting position, not the path of the next year.
Historical precision is limited. The long daily record rests on one full-history provider, early observations include predecessor-index history, and some early trading-session coverage remains uncertain. A second independently validated daily archive would strengthen the event counts. Overlapping samples, clustered crises and changing market structure also limit what historical frequencies can establish.
A verified year-ahead implied-volatility measure is unavailable, and near-term VIX cannot substitute for it. Better evidence on earnings revisions, borrower refinancing needs and the realization of large financing commitments would help distinguish temporary turbulence from a sustained stress regime. These gaps leave uncertainty around the forecast; they do not change the saved probabilities.
Forecast history
| Date | Forecast |
|---|---|
| 32% chance of yes |
Sources Prolepsia read
- S&P 500 (SP500) | FRED | St. Louis Fedfred.stlouisfed.org
- S&P 500 (^GSPC) Historical Data - Yahoo Financefinance.yahoo.com
- daily-price historyquery1.finance.yahoo.com
- FactSet Earnings Insightinsight.factset.com
- VIX Volatility Products | Cboewww.cboe.com
- Employment Situation Summary - 2026 M10 Resultswww.bls.gov
- Financial Stability Reportwww.federalreserve.gov
- Q2 2026 market outlook and positioninginstitutional.fidelity.com
- S&P 500® | S&P Dow Jones Indiceswww.spglobal.com
The question
This question asks whether the S&P 500 Index (SPX), the benchmark index of large-cap U.S. equities, will experience a peak-to-trough decline of strictly more than 15% at any point during the 2027 calendar year, from January 1 through December 31, 2027. The decline will be measured using official daily closing values of the S&P 500 price index, excluding dividends, rather than intraday highs and lows or total returns. A peak is defined as the highest daily closing level recorded from January 1, 2027, through the trading day immediately preceding the date being evaluated. A qualifying decline occurs when a subsequent daily closing value is more than 15% below that preceding peak. Only peaks and troughs occurring within 2027 will count; declines from peaks established before January 1, 2027, will not qualify unless the index establishes a qualifying peak during 2027 itself. As of November 15, 2026, the S&P 500 closed at approximately 7,561.67 on November 13, 2026, roughly 3.3% below its October 6, 2026, record closing high of 7,818.93. The index has experienced recent volatility but remains relatively close to its record high. The question therefore concerns whether a substantially larger market decline will occur during the following calendar year. Background sources include recent market reporting from the Associated Press and historical closing-level data. ([apnews.com](https://apnews.com/article/12c1e8155fb85c3c7b09e058ed6d3c74))
How it resolves
The question resolves YES if, on any regular U.S. stock market trading day between January 1 and December 31, 2027, inclusive, the official closing level of the S&P 500 price index is strictly more than 15% below the highest official closing level previously recorded on a trading day within that same calendar year. For each trading day t in 2027, define P(t) as the S&P 500 official closing level on that day and H(t) as the maximum official closing level recorded from the first trading day of 2027 through the trading day immediately preceding t. The drawdown percentage is calculated as 100 × (H(t) − P(t)) / H(t). The question resolves YES if this value exceeds 15.000% on any eligible trading day. For example, if the index reaches a closing peak of 8,000 during 2027 and subsequently closes below 6,800 during 2027, the question resolves YES. A closing level of exactly 6,800 would not qualify because the decline would be exactly 15%, rather than more than 15%. The question resolves NO if no qualifying decline occurs during 2027. The primary resolution source will be the official S&P 500 daily closing index levels published by S&P Dow Jones Indices (S&P DJI), using the standard S&P 500 price-return index (SPX), rather than the total-return or equal-weight index. If official historical closing data cannot be obtained directly from S&P DJI, historical daily closing data published by reputable financial data providers, such as Bloomberg, Reuters, or Yahoo Finance, may be used, provided that the data correspond to the standard S&P 500 price index. In the event of discrepancies, official S&P DJI data will take precedence. The question will be evaluated after the final scheduled U.S. equity trading session of 2027, using the official closing values available by January 15, 2028. If the necessary data remain unavailable or irreconcilably inconsistent by that date, the question will be annulled.
Fine print
The measurement is based exclusively on daily closing index levels, not intraday prices. A decline exceeding 15% during a trading session that subsequently recovers to 15% or less by the official close does not qualify. A decline must occur after a qualifying peak, and both the peak and the subsequent lower close must occur during 2027. A decline in early 2027 relative to a peak established in 2026 does not count unless a qualifying 2027 peak also exists. The reference peak is a running year-to-date closing maximum and does not need to be an all-time record high. Market holidays and weekends are excluded because there are no official closing values on those days. The index need not remain more than 15% below its peak for any minimum duration; one qualifying daily close is sufficient. The index may recover completely before December 31, 2027, without affecting a YES resolution. Dividend payments are excluded, and the performance of S&P 500 ETFs such as SPY or VOO is not used as a substitute for the index itself. Officially corrected index closing values will be used when available before the resolution deadline. If trading is suspended for part of the year, the question will still resolve using the official closing levels that are available, provided they are sufficient to determine the outcome. If the index is discontinued or materially redefined during 2027 and no sufficiently comparable official series is available to evaluate the full year, the question will be annulled.
More public forecasts
Will NOAA's Climate Prediction Center report a three-month Niño 3.4 index value (ONI or relative ONI, whichever CPC uses as its official index) for any season from SON 2026 to MAM 2027 higher than the peak of the 2015-16 El Niño on the same index?
84%chance of yes
Prolepsia forecast
Will the spread between French and German 10-year government bond yields close above 200 basis points on any day between 1 January and 31 December 2027?
43%chance of yes
Prolepsia forecast
Will OpenAI report that ChatGPT has 1.5 billion weekly active users at any point in 2027?
86%chance of yes
Prolepsia forecast
Will the gap between 10-year and 2-year US Treasury yields be above 75 basis points on 31 December 2027?
30%chance of yes
Prolepsia forecast
What will the spread between French and German 10-year government bond yields be at the close on 31 December 2027, in basis points?
143median
Prolepsia forecast