S&P 500 Historically: When Are All-Time High Streaks Exhausted?

Summary

• Analyzing historical data suggests that even when always buying at all-time-highs, attractive returns could be generated with the S&P 500.

• For long-term investors, history therefore shows that all-time highs are no reason not to buy the index.

• However, buying at all-time highs could indicate weak performance in the short term.

• Moreover, extended streaks of all-time highs indicate psychologically daunting temporary draw-downs ahead.

• I typically rate Hold/neutral, even when my inclination is bullish or bearish. Rating systems don't consider time horizons or investment strategies. My articles aim to inform, not to make decisions.

Thesis

The S&P 500 (SP500, SPX) is currently on a 6-month+ streak of reaching week-closing all-time highs when allowing for some weeks of tolerance with interruptions. This naturally opens discussions about whether buying at all-time highs is still sensible or, more specifically for me, how long such a streak could historically last. With this research article, I emphasize once again the well-known fact that buying at all-time highs has historically still promised attractive returns over the long term. Nevertheless, it obviously makes a difference whether it is the first or the last in a series of ATHs. The current series has not necessarily reached extremes yet, but already carries the potential for significant interim draw-downs.

Research Design

• Weekly S&P 500 index-closing data from 1970 to 2024 YTD.

• Therefore, an all-time-high (ATH) in this context is defined as a week-closing ATH.

• This research observes historical S&P 500 data in an isolated manner, i.e., without paying attention to other highly important factors like valuation, overall state of economy, monetary policy or index-constituents. These are clear limitations of this research article.

• Some observed time-periods were chosen arbitrarily to land at exemplary insights.

• While applying the greatest care in data analysis, this article is nonetheless not meant to be considered a comprehensive scientific paper.

• Keeping all of the above in mind, this research contributes a small excerpt to the abundance of related existing research and, with its limitations, opens up possibilities for future further research.

Investing At All-Time Highs

Firstly, with my research I want to add to the widely discussed common opinion among opponents of market-timing, claiming that one should invest at all times. This is because historically, even the worst market-timer has always yielded results that for some do not significantly differ to the worst compared to random investors and best market-timers. However, while this might be the case, and I am not at all about perfect market timing, I am into avoiding high draw-downs shortly after investing from a purely psychological perspective.

I have observed the performance of the S&P 500 within various periods following week-closing all-time highs, and have then annualized them for comparison. It indicates that the average annualized return within 1 week after a week-closing ATH was a lackluster 3.4%. The average annualized return within 4 weeks after a week-closing ATH stood at 5.6%, after 8 weeks 8.2%, and all the way up to 9.9% over a 3-year period after a corresponding ATH.

This indicates that all-time highs were, in fact, potential indicators for lackluster performance in the very short term of the following approximately 1 to 8 weeks, meaning buying at the ATH and selling 1 to 8 weeks later. However, on average, if always buying at an ATH and selling 6 months to 3 years later, the annualized return still stood at an attractive 8 to 10%.

All-Time-High Streaks

What, from my perspective, is lacking in all of the above is that no distinction is made between whether an all-time high was the first in a series of ATHs within a bull market, or the last one. Because, purely from intuition, the first ATH after a long period is more likely to be an indicator of further rising prices, which logically leads to additional ATHs, while after a series of ATHs there have typically been significant corrections.

First, I defined an ATH streak not only as the number of weeks that closed at an ATH, but also claim that one can designate a series as an ATH streak even when it is briefly interrupted by a few weeks without an ATH. For example, an ATH streak for me could be:

+ a 10-week series of ATHs,

+ followed by 4 weeks without an ATH,

+ and then another 3-week series of ATHs,

X and then I observed, let's say, 23 weeks without a new ATH.

In that case, my observed streak would total 17 weeks, given a 4-week tolerance.

I have observed week-streaks with 4 weeks of minimum tolerance, as well as 10 and 20 weeks of tolerance to capture long-term ATH streaks. Let's start with the latter – very long-term upward ATH streaks, with a tolerance of 20-week interruptions without ATHs.

The shortest ATH streak of that kind lasted 5 weeks in 2018. The longest streak lasted 268 weeks, over 5 years, from 1995 to 2000. Currently, we are in a 30-week streak that has been ongoing since the end of 2023 up to the present. This is approximately in the middle of all observations. Therefore, the historical probability that the current streak will continue is about 50:50, assuming a 20-week tolerance is allowed to define it as a streak.

But what happened after streaks that lasted longer than the current one? To address this, I investigated the maximum drawdown after the end of longer streaks until the next ATH. The table on the right should be read as follows:

The longest streak of 268 weeks ended with an index value of 1528 at the beginning of 2000, resulting in a maximum drawdown of -48%, and it took 7 years until the next ATH, which occurred in 2007. On average, the maximum drawdown after a streak that lasted longer than the current one was -23%.

Let's now look at a somewhat shorter time frame and define an ATH streak only if it is not interrupted by more than 10 or 4 weeks. Naturally, with fewer weeks of tolerance to call it an ATH streak, the number of observations increases, while the duration of each streak decreases. The longest streak with a 10-week tolerance lasted 121 weeks from 2013 to 2015. The current streak since the end of 2023 is in the top 30% of the longest streaks with its 30 weeks. Even with just a 4-week tolerance, we are currently already in the top 30% of the longest streaks. The current streak has been ongoing since May and has lasted 10 weeks, with the longest such streak being 44 weeks from late 2020 to mid-2021.

Takeaway

In summary, while we are not yet among the top values for long-term ATH streaks (with a 20-week tolerance), the maximum drawdown during streaks that lasted longer than the current one averaged -23% and ranged up to -48%. In the shorter term, with less tolerance, we are in the upper third of the longest ATH streaks, indicating that while the streak could naturally continue, we are already approaching the longest historical ATH runs in the short term.

Combined with the insight that even the worst market timers have achieved attractive returns over the long term, I, personally, would continue to buy the index but at a significantly reduced rate currently, as I consider a drawdown highly likely — though I am like any other investor unable to time it optimally. Therefore, my cautious hold rating should be taken quite literally.