Can anyone predict a stock market crash? The track record of forecasters

What counts as a correct call when forecasters predict a stock market crash?

A correct call states a size of fall, a window, and a reason before the event, in writing and with a date. A fall of 20% in a year, announced in advance, counts. A comment that markets look expensive does not, because any outcome fits it. The standard sounds strict, yet it is the only one that lets you compare 2 forecasters fairly.

Forecasters also differ in what they claim. Some describe conditions, some give probabilities, and some call dates. Grading each claim type on its own scale avoids a common error, where a good reading of conditions gets blamed for a missing date, or a lucky date gets praised as a method.

Which scorecard measures a forecaster who calls stock market crashes?

Four numbers describe a forecaster: the hit rate, the false alarm rate, the lead time, and the size of the fall called against the fall observed. Imagine a forecaster who issues 10 warnings and sees 2 crashes follow. The hit rate looks like 20%, and the cost of the 8 false alarms matters as much as the 2 hits, since each one asked someone to act.

Lead time deserves its own column. A warning that arrives 3 years early forces a long wait, and a position held through a wait has a cost. The most useful forecasters publish all their calls, wrong ones included, so anyone can fill in the scorecard from the archive. A public archive is the best credential.

Why do lucky forecasters look like experts after a stock market crash?

After every large fall, headlines celebrate the few voices that warned in advance. Among thousands of commentators, some warn at any time, so a handful will match a crash by chance. The crowd that warned and saw the market keep rising is forgotten. This survivor effect makes forecasters look better than the full record supports, and it flatters the loudest voices most.

A fair test looks at the whole group instead of the winners. Count everyone who made a similar call in the same period, then ask how many were right. If the winners are a small share and only the outcome separates them from the rest, luck explains the result as well as skill does. The page on who predicted the 2008 crash applies that test.

Did DSGE models predict the 2008 financial crisis in advance?

DSGE models, the main tool at the IMF, the Fed, and the ECB, returned no systemic risk before 2008, according to the EQOS Asia 2008 back-test (2026). The study explains why: the models assume rational agents, perfect information, and linear behaviour, so ruptures cannot appear by construction. A tool built that way cannot flag a crash, whatever data you feed it.

Public statements of the period reflected the same blind spot. The study quotes Greenspan saying that the market regulates itself, and Bernanke in March 2007 saying that the subprime impact would be contained. A forecaster who relies on such premises inherits their limits, which is why the premises deserve a check before any forecast does.

Can a measured score beat a stock market crash forecast?

A measured score offers a different route: instead of forecasting a date, it reads the structure of the system. According to the EQOS Asia 2008 back-test (2026), public data from 2006 and 2007 gives the global financial system a φ(S) of 0.234, below the 0.3 critical threshold, with 6 of 9 dimensions in the critical zone. The data came from the SEC, the BIS, the FDIC, and Case-Shiller.

The study also places the point of no return about 1.1 years after mid-2006. Treat this with care: it is a retrospective exercise, run in 2026 with knowledge of the outcome, and the study calls itself an educational analysis. Its value is the method. Each input is public, each step can be recomputed, and 2 readers can trace a disagreement to a specific number.

What limits any forecaster who tries to predict a stock market crash?

Triggers arrive from outside the data. On 19 October 1987 the Dow Jones Industrial Average fell 22.6 percent in one session, a record documented by Federal Reserve History. Fragility had built for months, and the exact day still came as news. Conditions give the size of a possible fall, and the trigger sets the calendar. Forecasters who separate the 2 stay credible.

Crowd behaviour adds a second limit. A 2011 study from the New England Complex Systems Institute, summarised on Wikipedia, found a sharp rise in investor imitation in the year before each market crash it examined, which the summary calls an advance warning sign. Such research points to behaviour that can be watched, yet it leaves the date open.

How should you read a stock market crash call from a forecaster today?

Ask 4 questions of any call. Does it state a size, a window, and a reason? Does the forecaster publish earlier calls, misses included? Does the reasoning rest on public numbers that you can recompute? Does it offer a probability and a review date? A call that answers all 4 deserves a place in your dashboard, and the rest stays as background reading.

A 30-minute pre-diagnostic with EQOS Asia applies the same scoring to your own exposure, and the Λ simulator lets you test the outcome in a soft landing, a correction, and a crash. This page is general information, not investment advice. As of October 2026 the habit that works is to score forecasts before acting on them.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile