What are the signs of a market crash? (And why most predictions fail)

This page is educational and does not constitute personalized investment, tax, or legal advice. Clockwise Capital is a registered investment adviser. The indicators discussed below are commonly used analytical tools — none provides a reliable forecast of timing, and decisions about your portfolio should be made in the context of your full financial picture.

There is a long list of things people will tell you mean a crash is coming. Some of them are useful diagnostic tools. None of them, individually or in combination, has a track record reliable enough to trade on. This page walks through the main ones honestly — what each measures, what its historical signal-to-noise ratio looks like, and why the most actionable response is usually the same regardless of what the indicators say.

The inverted yield curve has the cleanest historical record — but its timing is unusable

When short-term Treasury yields exceed long-term yields, the yield curve is inverted. This is unusual — investors normally demand higher yields for tying up money longer — and it has preceded most U.S. recessions of the past 50 years. The 2-year-to-10-year spread is the most-watched version, though some economists prefer the 3-month-to-10-year. Federal Reserve research and the New York Fed's recession probability model both treat curve inversion as a meaningful signal.

What's also true:

• The lag between inversion and the start of a recession has ranged from roughly 6 months to over 2 years.
• A recession is not the same as a stock market crash. Equity markets sometimes peak before recessions, sometimes during, and sometimes recover well before the recession ends. The 2020 episode illustrates the disconnect — the recession was severe, the equity drawdown was deep, but the recovery was much faster than recession standards would suggest.
• Recent inversions have produced unusual outcomes — in particular, an inversion that lasted longer than any prior episode without an obvious recession at the expected time. Whether this means the indicator has weakened, or the recession has merely been delayed, is genuinely debated.

The honest read: the yield curve is a real signal that something stressful in the credit/economic system tends to follow. It is not a tool you can use to enter or exit equities with confidence.

CAPE and other valuation indicators show "eventually" but not "when"

Yale economist Robert Shiller's cyclically-adjusted price-to-earnings ratio (CAPE) divides current price by inflation-adjusted average earnings over the prior 10 years. The smoothing reduces noise from short-term earnings cycles. Long-run data — Shiller maintains a public dataset — shows that high CAPE readings have been associated with lower forward 10-year equity returns.

The frustration with CAPE as a timing tool:

• It has been above its long-run average for most of the past 25 years, including during periods of strong returns.
• The "right" comparison level has shifted. Some researchers argue lower interest rates structurally support higher CAPE; others note that the composition of the index has changed (more software, less industrial). Either way, "above historical average" has not been a useful sell signal.
• The relationship is statistical and decadal. Even when CAPE has correctly predicted lower returns, "lower returns" can mean a slow grind, not a crash.

The Buffett Indicator (total U.S. market capitalization divided by GDP) tells a similar story. Useful as a long-run valuation context. Not useful as a market-timing tool, especially given the increasing share of S&P 500 revenue earned outside the U.S.

Sentiment and positioning indicators are noisy contra-indicators

A long list of measures purport to capture investor mood — the AAII Investor Sentiment Survey, the CNN Fear & Greed Index, the Conference Board's confidence indexes, options-market positioning, and inflows/outflows from various ETFs. The classic interpretation is contrarian: extreme bullishness precedes weakness, extreme bearishness precedes strength.

What the data actually supports:

• At extreme readings, sentiment indicators have some contra-indicator value — historically, the worst times to invest have come after periods of widespread euphoria, and the best times have followed periods of widespread despair.
• Most of the time, sentiment is in the middle and tells you nothing actionable.
• "Extreme" is itself a moving definition. The euphoria of 1999 and the euphoria of 2021 looked similar but produced very different downside paths. Sentiment is a confirmation tool, not a primary indicator.

Margin debt and credit conditions describe fragility, not timing

FINRA publishes monthly aggregate margin debt — the total amount investors have borrowed against their securities to buy more securities. Rising margin balances tend to accompany late-cycle markets, because investors get more comfortable using leverage when things have been going well.

The use case for this indicator is different from the others. It does not say "a crash is imminent." It says "the conditions for a sharp decline — should something trigger one — are present." Forced selling driven by margin calls is one of the mechanisms by which corrections turn into crashes. So margin debt tells you about the fuel, not the spark.

Credit-spread widening (the difference between corporate bond yields and Treasury yields) is a related fragility signal. The Federal Reserve tracks several variants. Like margin debt, widening spreads describe a system under stress without necessarily forecasting an exact moment of break.

Breadth and market internals tell you whether the rally is broad or narrow

When a few large stocks are driving most of the index's gains and the median stock is flat or declining, breadth is "narrow." This has often — though not always — preceded broader market weakness. The intuition is that narrow leadership reflects an economy supporting fewer healthy companies than the headline implies.

Breadth indicators include:

• The advance-decline line (cumulative count of stocks rising vs. falling each day)
• The percentage of stocks above their 50-day or 200-day moving average
• Equal-weight versus market-cap-weight performance gaps

These are diagnostic — they tell you what kind of market you are in. They are not a timing tool. Markets can stay narrow for extended periods, and not every narrow market produces a meaningful drawdown.

Why combining indicators does not solve the timing problem

A common move is to combine multiple indicators, on the theory that several agreeing reduces noise. The intuition is reasonable; the practical results are mixed. The same indicators that produce false positives individually often produce false positives together — they are correlated, because they are all measuring related things about late-cycle markets. Researchers who have backtested various indicator-combination strategies generally find modest improvements in long-run risk-adjusted returns at best, and significant transaction costs and tax drag in practice.

The honest takeaway from decades of quantitative research on market timing is the one DALBAR's Quantitative Analysis of Investor Behavior keeps surfacing: average investors who attempt to time the market significantly underperform investors who don't. The indicators are real; the ability to act on them well is rare.

So what should you actually do?

Indicators are diagnostic. They are useful for understanding what kind of market you are in and where pockets of fragility exist. They are not useful as timing tools. The practical response, regardless of what the indicators are saying, is the same set of actions covered in How to prepare for a market crash:

• Match your allocation to your actual risk capacity, not just appetite.
• Build enough cash and short-duration assets to fund near-term spending without selling equities at a low.
• Rebalance on a schedule.
• Write down a plan you will follow when headlines turn.
• Reduce check-frequency and add behavioral guardrails.

If indicators are flashing and you feel an urge to act, that is a signal to talk to someone — a fiduciary advisor, ideally — before doing anything irreversible. The cost of waiting 48 hours and getting a second opinion is almost always smaller than the cost of acting on emotion. For a related framework, see Should I sell my stocks now?.

The most reliable indicator is the one in the mirror

After a long career and a lot of cycles, most experienced advisors arrive at a similar conclusion: the indicator that best predicts whether an investor will do well over a full cycle is not the yield curve or CAPE. It is the investor's own behavior. The plan you actually follow beats the plan you intended to follow. Every framework on this page exists to help you build a plan that works regardless of what the next twelve months bring.