With dividends counted, SPY (the S&P 500 ETF) closed at a new high on 852 trading days since 27 January 1994. We call them high days. A year later, SPY stood higher after 83.9% of the 815 high days that had a full year ahead. After all non-high days, the share was 81.1%.
SPY is an exchange-traded fund, or ETF, that tracks the S&P 500. We compared buying on a high day with buying on any other day.
Is buying at all time highs a bad idea? An all time high is a close above every earlier close in our SPY data, with dividends counted. In SPY since 1994, buying on such a high day showed no clear gap in median returns against buying on any other day.
One year later the median return was 16.5% after a high day and 14.7% after any other day (not a high day). Three years later it was 35.1% against 40.9%.
A year later, 16.1% of high-day buys were lower, against 18.9% for other days.
The worst one-year result from a high day was a 36.4% loss.
What counts as an all time high?
Our SPY price file starts on 29 January 1993. We used its daily closing prices to 26 October 2026, 8,493 trading days. We call a day a high day when its close is above every earlier close in the file. A close that only equals the old high does not count.
We left out the first 252 trading days, because at the start of a price history nearly every day is a high. Counting starts on 27 January 1994. That leaves 8,241 days, of which 852 are high days, or 10.3%.
High days came in 24 of the 33 calendar years from 1994 to 2026. SPY set no new high in the other 9. The 852 high days fall into 52 stretches, separated by gaps of more than 21 trading days.
How did we compare high days with other days?
For each day we measured the return from that close to the close 21, 63, 126, 252 and 756 trading days later (the horizons). These are about 1 month, 3 months, 6 months, 1 year and 3 years.
Other days means every counted day that is not a high day. The median is the middle return once all are in order. The mean is their sum divided by how many there are.
The windows overlap, because a buy on Monday and a buy on Tuesday share almost the same year. The test holds far fewer than 815 independent outcomes.
What did SPY return after high days?
The table gives the median return and the share of buys that ended higher.
Horizon | High-day buys | Median, high days | Median, other days | Share higher, high days | Share higher, other days |
|---|---|---|---|---|---|
1 month | 851 | 1.1% | 1.5% | 65.9% | 65.6% |
3 months | 847 | 3.8% | 3.9% | 76.3% | 71.1% |
6 months | 834 | 7.5% | 7.1% | 81.5% | 75.6% |
1 year | 815 | 16.5% | 14.7% | 83.9% | 81.1% |
3 years | 713 | 35.1% | 40.9% | 87.5% | 82.6% |
SPY with dividends counted, 27 January 1994 to 26 October 2026. Share higher is the share of buys with a gain at the end of the horizon. Source: YX Insights price data
At one year, the median return was 16.5% after high days and 14.7% after other days. At six months it was 7.5% against 7.1%. At one month and three months the high-day medians were lower, 1.1% against 1.5% and 3.8% against 3.9%.
At three years the median was 35.1% after high days and 40.9% after other days. The mean ran the other way, 42.7% against 39.5%. Chart 1 shows the medians.
Source: YX Insights price data
The share of buys that were higher was greater after high days at all five horizons. At three months it was 76.3% against 71.1%. At three years it was 87.5% against 82.6%. At one month the shares were almost level, 65.9% and 65.6%.
How often did buying at a high lose money?
After one year, 131 of 815 high-day buys (16.1%) were lower. For other days, 1,354 of 7,174 (18.9%) were lower. After three years, 89 of 713 high-day buys (12.5%) were lower, against 1,177 of 6,772 (17.4%).
A buy can end higher and still fall on the way. We also found the lowest close in the 252 trading days after each buy. The median was a fall of 4.1% after high days and 5.1% after other days. A fall of 20% or more followed 9.8% of high days and 15.1% of other days.
Fewer losses do not make a high day safe. Among the 815 high days with a full year ahead, the worst one-year result was a 36.4% loss, from the 9 October 2007 close to the 8 October 2008 close.
Could chance explain the gap?
High days cluster in time, so we ran a check. We took the pattern of high days, clustering included. We then moved it round the calendar by a random number of days, with days pushed past the end returning to the start.
We did this 2,000 times. Each time we recomputed the gap at the five horizons: the median after the moved high days minus the median after the other days. That is five tests.
At every horizon, the real gap sat inside the middle 90% of the moved gaps. At one year the real gap was 1.9 percentage points, from the unrounded medians. A gap that large or larger came up in 15.8% of the moved patterns.
At three years the real gap was -5.8 percentage points. A gap that low or lower came up in 9.6% of the moved patterns. Chart 2 shows all five.
Source: YX Insights price data
The data show no detectable difference in median returns between high days and other days. They cannot show that none exists, because 33 years of one fund is a small sample.
Does the result depend on how we define a high?
We reran the comparison three more ways. We left out the first 126 days, then the first 504 days. Then we defined highs on SPY's quoted price close, without dividends, which gives 693 high days against 852. Returns stayed dividend-adjusted.
In all four versions the median was higher after high days at 6 months and 1 year. It was lower at 1 month, 3 months and 3 years. No median moved by more than 1.1 percentage points from the main version.
What happened after the four falls of 20% or more?
SPY closed 20% or more below its highest close to date in four stretches since 1993. The high before each fall was a high day by our rule. We measured each fall from the high to the lowest close before SPY next closed above that high.
The 24 March 2000 high fell 47.5%. SPY first closed above it 79.1 months later. The 9 October 2007 high fell 55.2%, with a first close above it after 58.3 months.
The 19 February 2020 high fell 33.7%, with a first close above it after 5.7 months. The 3 January 2022 high fell 24.5%, with one after 23.3 months.
Source: YX Insights price data
A buyer at those four highs was lower after one year in three cases: 2000, 2007 and 2022. The 2020 buyer was 17.7% higher. Our guide to stock market recovery time measures a recovery that holds, so its dates run later than the first close above a high. Its 48.5% fall for 2000 also runs to a later low.
What does the test leave out?
The test covers one fund over one 33-year history. Four falls of 20% or more are too few to say how a future fall would look. We leave out taxes, trading costs and the interest on cash.
The test does not try waiting for a fall, or spreading purchases over time. Our comparison of the lump sum against spreading money in covers that choice. Our guides to time in the market, the drawdown and how to read a backtest cover related ground.
In SPY since 1994, buying on a high day showed no clear gap in median returns against buying on other days. The four highs before large falls took 5.7 to 79.1 months to be exceeded.
Get the next explainer
The Academy is free to read, here on the web. The Long Trend, our free weekly letter, comes by email.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
Common questions about buying at all time highs
Is it bad to buy at an all time high?
Not in SPY's history, on our test. After one year, 83.9% of the 815 high-day buys were higher, against 81.1% for other days. Still, 131 of the 815 were lower. The worst lost 36.4%. The gap in median returns fit chance.
Do stocks keep rising after an all time high?
In SPY, usually: 83.9% of buys were higher after one year. The median one-year return from a high-day close was 16.5%. After three years the median was 35.1%, with 87.5% higher. After one year, 16.1% lost money. The largest one-year loss was 36.4%.
What happens if you buy at the top of the market?
A top is only known afterwards. Among the four highs followed by a fall of 20% or more, SPY first closed above the high 5.7 to 79.1 months later. Buying at the 2020 high was 17.7% higher a year later. Buying at the 2007 high was 36.4% lower.
How often does the stock market hit all time highs?
SPY closed at a new high, with dividends counted, on 10.3% of trading days from 27 January 1994 to 26 October 2026, 852 of 8,241 days. It set at least one high in 24 of the 33 calendar years. It set none in the other 9. High days arrive in clusters, not evenly.
DISCLAIMER: This article is intended for educational purposes only. Any information or analysis in this note does not constitute an offer to sell or a solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice, nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.
