All-Time High

An all-time high is the highest value a security, fund or index has reached at any point in its recorded history, usually quoted on closing prices. Reaching one means only that the series has exceeded every earlier reading; nothing about the level itself distinguishes a record from any other price. It is the mirror image of a drawdown, which is measured downward from exactly this reference point, so an investment at an all-time high is by definition an investment with no current drawdown.

The phrase is often used interchangeably with "record high" and is frequently confused with a 52-week high, which is a different and much weaker statement: a 52-week high is only the highest level of the past year, so a security can set one while remaining far below its all-time high.

Two things make records less remarkable than the coverage of them suggests. The first is arithmetic. Once a series has climbed past its previous peak, every subsequent step up is a new record, so records are not scattered evenly through history. They arrive in runs, separated by long stretches in which none occur because the market is below an earlier peak. A count of record days therefore says more about which regime the market is in than about how unusual any individual record is.

The second is that "the level" is not one number. A price index tracks prices only and leaves out dividends; a total-return index assumes dividends are reinvested and so is always higher over long periods, and reaches its records earlier. An inflation-adjusted series is different again, because prices have to rise faster than the cost of living for a real record to be set. The headline figure quoted in the press is almost always the nominal price index, which is the one that sets records most rarely of the three and the one that corresponds least well to what a diversified investor actually holds.

Individual investors have a fourth version that matters more than any of them, which is their own account's high-water mark. That figure depends on the contributions and withdrawals made along the way, so an account can be at its own all-time high while every index it holds is below theirs, simply because money kept arriving. The reverse also happens in retirement, when withdrawals can keep a balance below its old peak through a market that has fully recovered.

The behavioral question, which is why this page carries the psychology topic, is why a record level feels like a bad moment to invest when the same money invested a week earlier at a slightly lower price felt fine. The mechanism is reference-dependence: the old peak becomes an anchor, and prices above it read as expensive relative to that anchor rather than as ordinary. Our pages on anchoring bias and recency bias cover the mechanism. What this page can say plainly is the arithmetic point: in a series that rises over long periods records recur by construction, so a rule of "wait until it is not at a high" has no stopping condition other than a decline, which makes it a rule about the calendar rather than about value.

Take the running maximum of a series. Whenever the current value exceeds it, a new all-time high has been set and the running maximum resets to the new level. Everything between two records is a drawdown from the earlier one.

A hypothetical illustration of why the measure being used decides the answer. An index closes at 5,000 in one year, falls, and six years later closes at 5,200. In nominal terms that is a record: 5,200 is higher than any earlier close, and it will be reported as one.

Now suppose consumer prices rose 20 percent over those six years. In the earlier year's dollars the new level is 5,200 divided by 1.20, which is about 4,333. Measured that way the index is roughly 13 percent below where it stood six years earlier, and no record has been set at all. Both statements are correct descriptions of the same closing price.

A third measure would change it again. If the index's constituents paid dividends over those six years, the total-return version of the same index would have passed its old peak well before the price version did, because the reinvested income compounds on top of the price change.

What the term is useful for

• It is a precise, unambiguous reference point, unlike most market labels: the highest close ever recorded is a fact rather than a convention.
• It defines the zero point for measuring drawdowns, which is how portfolio declines are quoted.
• Tracking how far an account sits below its own high-water mark is a more useful personal statistic than tracking an index's records.

What it is not

• It carries no information about future returns, and it is routinely reported as though it does.
• It is normally quoted on a nominal price basis, which excludes dividends and ignores inflation, so it understates how often a diversified investor's holdings have actually made new highs.
• It invites a waiting rule that has no stopping condition, since there is no level at which a market becomes "not at a high" other than by falling.
• It is easily confused with a 52-week high, which is a far weaker claim.