Market declines get measured by depth. A fall of 10% is a correction, 20% makes it a bear market, and coverage tends to stop there.
Depth turns out to be a poor predictor of what happens next. Two declines of similar magnitude can recover in a year or take a decade, and the difference has almost nothing to do with how far prices fell.
What determines it is the cause. There’s an established framework for classifying declines by trigger rather than size, and the recovery profiles attached to each category differ enormously.
Where a Correction Sits on the Scale
Asking what is a stock market correction produces the 10% threshold, which describes magnitude and nothing else.
The threshold definitions are useful for communication and limited for analysis:
- Pullback, roughly 5%, common enough to be unremarkable
- Correction, 10% or more from a recent peak
- Bear market, 20% or more, conventionally over at least two months
- Crash, describing speed rather than depth
None of these say anything about duration, cause or recovery. A classification by trigger does.
Three Types of Decline
The widely used framework identifies three categories, each with different characteristics.
Research setting out the framework describes structural declines as the worst type but fortunately the rarest, always preceded by asset bubbles usually backed by high private sector debt, where the burst triggers a deleveraging cycle and often a problem in the banking or real estate sector, with equities falling 50% to 60% over three or four years and typically taking about a decade to recover fully in nominal terms.
The same framing describes cyclical declines as oscillating around economic cycles, where prices fall to reflect an anticipated downturn in activity and profits and recover once a policy response improves the outlook. Event-driven declines are triggered by a shock that derails an economic cycle, such as a war or a commodity crisis.
The critical distinction it draws: cyclical and event-driven declines don’t differ much in depth, typically 25% to 30%, but differ substantially in the speed of decline and recovery. The main determinant is whether a recession follows.
The Numbers Behind Each Type
Analysis of the historical record puts figures on those differences.
One examination of the data found that the average decline in event-driven bear markets of 29% is similar to the average of 31% for cyclical ones, but while cyclical declines lasted an average of two years and took four years to recover, event-driven declines lasted an average of just nine months and had recovered to previous peaks after just over a year.
The same analysis notes that volatility during the decline differs by type, averaging 14% of monthly returns from peak to trough in cyclical declines against 20% in structural ones, and that corporate profits typically begin recovering around ten months after a cyclical decline ends.
Nearly identical depth, four times the recovery period. That gap is the single most useful thing in the framework.
Why Identification Is Hard in Real Time
The obvious application is to identify the type early and respond accordingly. The research is candid that this is difficult.
Most structural declines begin looking cyclical. Many event-driven declines start with a specific shock, then raise recession risk and morph into something cyclical. Classification is straightforward in retrospect and genuinely uncertain while it’s happening.
That limitation is worth stating plainly, because a framework that only works after the fact can’t be used for timing. What it can do is set expectations, which is a different and more achievable goal.
What This Changes for a Holder
- Expect wide variation in recovery periods, rather than assuming an average
- Watch whether recession risk is rising, since that’s the variable separating the two common types
- Treat depth as uninformative about how long a recovery will take
- Size a cash buffer for the longer scenarios, not the median one
- Avoid assuming rapid recoveries, since the most recent examples have skewed expectations
The fifth point matters most for anyone whose investing experience began recently. Several recent declines recovered unusually quickly, which is not the historical norm across all types.
What the Taxonomy Doesn’t Do
It doesn’t predict declines, identify them reliably in real time, or tell anyone when to act. Applied carelessly it invites exactly the kind of forecasting the evidence doesn’t support.
Its value is in expectation-setting. An investor who knows that a decline of a given depth might resolve in a year or in a decade, depending on a cause that won’t be clear for months, plans differently from one who assumes recoveries take a predictable period.
That planning shows up in practical places. How much cash to hold, what withdrawal flexibility to build in, and how long a horizon a given holding genuinely needs before its exposure is appropriate.
