All-Time Highs

There's a hullabaloo recently that the S&P 500 (and nearly all market measures) have made all-time highs (or remade, to be more accurate). What's missing here is the glide path and a reality check. Here's SPY, the S&P 500 tracking ETF, for year-to-date 2025 and since inception as of this writing:

SPY = S&P500 SPDR
Max = All-Time High

Looking at the year-to-date chart one thinks: "not so much". Looking at the long-term chart, the market looks like its ALWAYS at all-time highs! This still begs the question: do you dare buy stocks at all-time highs? Or do you wait? First lets see how rare are all-time highs? And, do they matter? Here's the chart of the yearly percentage of all-time highs since the inception of SPY on 2/2/1993 and each year's return:

Annual Return = Continuously compounded annual return.

% ATH = Percentage of year's trading days at all time highs. 

Closing price source: Barchart.com

For instance, in 1995, SPY was at all-time highs for 23% of all trading days and closed up 32.1%. In 2008, SPY had NO, ZERO, NADA all-time highs and ended down 45.9%! Technically, there's a .56 correlation coefficient between all-time highs and annual returns. In practice, the ten years of zero all-time highs skew the results. There probably IS a stronger correlation of market highs and market returns. 

As to the burning question, is it good or bad to buy at all-time highs? To measure this, I look at the rolling returns for buying SPY on all days, days in years with all-time highs (normal years, excludes the 10 years with no all-time highs) and only on all time highs. Here we go:

For instance, the average 1 day return for every day SPY has traded is 0.04%. Excluding years with no all time highs rises the daily return to 0.06%. Buying on all-time high days raises your average 1 day return to a fantastic 0.59%. This is not uniform across holding periods. Holding periods greater than 5 year's have little effect. The 1 year holding period buying at all-time highs is dramatically higher, at 21.47%, versus buying just any day or even excluding days with no all-time highs. With the above results, one can conclude that buying at all time highs has HIGHER returns than buying on other days. It's almost an obvious result. 

It's hard to know if a year is "normal" in this definition of "normal=year with all-time highs". You may not know until the last day of the year. It's easy to buy at all-time highs. You just have to wait. Sometime you have to wait for years!