When the stock market hits a new record, many investors hesitate. Buying "at the top" feels like the riskiest possible moment. But record highs are far more common than most people think — and what happened after them is more nuanced than either "never buy at the top" or "highs don't matter".
We looked at every month since 1871 in which the S&P 500 set a new all-time high, and compared what happened next with all other months.
Record highs are normal
Using monthly average prices, the S&P 500 set a new all-time high in 335 of the 1,869 months from 1871 to September 2026 — about 18% of the time. Since 1950 the figure is 28%: more than one month in four. Between January 2024 and September 2026, the index set new monthly highs in 23 of 33 months.
A market at a record high is not unusual. It is what a market that rises over the long run looks like most of the time.
What happened next
The table compares annualised total returns (dividends reinvested) after record-high months with returns after every month in the dataset:
| Holding period | Median return after a new high | Median return, all months | Chance of a loss after a new high | Chance of a loss, all months |
|---|---|---|---|---|
| 1 year | 11.9% | 10.8% | 26.5% | 27.8% |
| 3 years | 10.0% a year | 9.9% | 15.9% | 15.4% |
| 5 years | 9.6% a year | 9.6% | 20.6% | 10.5% |
| 10 years | 7.4% a year | 8.7% | 10.2% | 3.0% |
Two things are true at once.
The typical outcome barely changed. Over one, three and five years, the median return after a record high was almost identical to the median return from any other month. If you buy at a new high, the most likely result is roughly the same as buying any other time.
The bad outcomes were more common over longer periods. Over five and ten years, the chance of a loss after a record high was noticeably higher than average. That is mostly because of two periods: the late 1920s and the late 1990s. In both, the market set many highs in a row before a historic crash, and an investor who bought at those peaks waited a very long time. The longest gap between highs in the whole dataset ran from September 1929 to September 1954 — 25 years.
Since 1950
Restricting the data to 1950 onward gives a similar picture. The median one-year return after a new high was 11.4%, against 13.4% for all months, and the chance of a one-year loss was 26% against 22%. Over five years, the median was 10.6% a year after a high and 11.7% for all months.
So new highs have not been a reliable signal to wait. But they also have not been a free pass.
Why "waiting for a dip" usually doesn't help
If highs are common and the market tends to rise, waiting for a pullback often means waiting while the market moves further away. Our lump sum vs dollar-cost averaging analysis found that investing immediately beat spreading the money over 12 months in about seven out of ten historical periods — and many of those periods began near a high.
When to be more careful
The weaker long-run results after highs were concentrated in periods when prices had run far ahead of earnings. A record high with a moderate valuation is very different from a record high at extreme valuations like 1929 or 1999. Our guide to CAPE and future returns shows how strongly starting valuations have been linked to the following decade's returns.
Practical takeaways
- Don't treat a record high as a reason to stay out. Historically, the typical outcome after a high was about the same as at any other time.
- Do size your risk for the bad case. Because the bad outcomes after highs were worse, invest only an amount you could leave alone through a long drawdown. See how long bear markets took to recover.
- If you're anxious, average in over a short period. Spreading a lump sum over three to six months gives up a little expected return in exchange for less regret if a fall follows.
- Keep investing regularly. Monthly contributions automatically buy at highs, lows and everything in between. Try different contribution levels in our calculator.
Notes
Data: Robert J. Shiller, U.S. Stock Markets 1871–Present, via datasets/s-and-p-500. A "new high" is a month whose average price exceeded every previous monthly average. Forward returns use total returns with dividends reinvested, available through June 2023; price data after that comes from FRED. Monthly averages smooth out daily swings. Calculations are our own.
This guide is educational and not investment advice. Past performance does not guarantee future results.