Too Much of a Good Thing

In the most recent two months, however, a small group of stocks has allowed the S&P 500 to reach the upper bounds of this range, while an equal-weight version of the index has struggled. As of the date of this report, 50.8% of the S&P 500 had year-to-date losses ranging from as little as a fraction of a percent to as much as 52%.

The recent dramatic outperformance of a small subset of the market has drawn intense interest, particularly after the same groups of stocks suffered through a difficult 2022 and the early part of this year. On average, the stocks listed in Chart 1 as a group, year-to-date as of May 18, 2023, had an average gain of 55.62%. They are only 1.59% of the number of issues in the S&P 500, but they are 25.93% of the market value of the index.

This has led to the capitalization-weighted S&P 500 having a year-to-date gain 6.6 times more than the equal-weight S&P 500 Exchange Traded Fund (ETF), while the tech-heavy Nasdaq Composite had a gain 2.27 times greater than the cap-weighted S&P 500 and 15 times more than the equal-weight S&P 500 ETF.

Disparities like these are both extreme and usually unsustainable, but they resulted from investors seeking growth potential as the overall economy languished. Interest in artificial intelligence (AI) exploding added to the buying rush.

Valuations for many of the stocks that have significantly appreciated recently are relatively high, but dramatically lower than those in the late-1990s tech bubble. However, the biggest U.S. stocks are up significantly more than the rest of the market this year and trade at a 30 price-earnings ratio compared to approximately 17 for rest of the S&P 500.

The current AI mania that sparked much of the recent activity is reminiscent of the internet in its earliest development, when even the remotest connection to the internet generated gains in related stocks. Undoubtedly, AI has enormous potential—possible even much more than the market is now considering, but even the most significant ideas often take many years to develop. It took decades, for example, for the internet to be widely adopted.

Many of the stocks that are thought to be involved in AI development are significantly stretched, while some do not deserve to even be in the conversation.

Caveat emptor is appropriate now.