America’s most famous stock-market measure is more broken than usual

Far from tracking the S&P 500 index of the top U.S. companies, the Dow Jones Industrial Average for the past two years has been left in the shade. It lagged behind the S&P by more than 10 percentage points in 2023 and 2024, a dismal performance matched in only two years since the S&P 500 was introduced in 1957. (Both, 1980 and 1998, were times of economic turmoil.)

The Dow’s current problem is Big Tech. But the reason it has this problem is more fundamental: The Dow is an awful index that was designed for an era of human stock selection and slide rules and is horribly outdated in the automated age of index tracking.

Professional investors mostly ignore the Dow, though they pay close attention to the Nikkei 225, Japan’s equally flawed index. Yet the Dow gets outsize attention in the media and among ordinary investors, presenting a distorted picture of what’s going on with America’s stocks.

Start with the current problem. Excitement about artificial intelligence has driven the Magnificent Seven stocks of Alphabet, Amazon.com, Apple, Meta Platforms, Microsoft, Nvidia and Tesla to make up almost a third of the S&P 500. Moves in the S&P are dominated by whether it is a good day or a bad day for Big Tech. But in the Dow, these seven make up only 13.9%, with Meta, Tesla and Alphabet missing altogether.

Instead, the Dow gives five stocks that barely matter to the market as a whole the same weight that the Mag 7 have in the S&P. Aside from Microsoft (third-biggest in the Dow, as in the S&P), the five-biggest Dow stocks are Goldman Sachs, UnitedHealth, Home Depot, Caterpillar and Sherwin-Williams. Together they are 32% of the Dow, but make up a tiny 2.6% of the S&P.

What moves the S&P matters relatively little to the Dow, and increasingly what moves the Dow matters even less to the S&P.

The reason for all this is a design decision that is bizarre today but made sense when the Dow was first compiled: It is weighted by price, not by market value. The higher the share price, the more a stock moves the Dow.

This is even though the share price matters not at all in the real world. Moves in the price matter of course, but whether a stock is say, $62, like Cisco Systems, or $647, like Goldman, makes no difference. Price-weighting made it easier to calculate the index in the days before computers and easy access to corporate share counts, but it is an anachronism today.

On top of that, the Dow isn’t made up of the biggest companies, but is supposed to be representative of the U.S. An index committee at S&P Dow Jones Indices, part of S&P Global, decides which companies get into the index. The committee—which includes two Wall Street Journal editors—has often picked new members at exactly the wrong moment, including the addition of Microsoft and Intel shortly before the dot-com bubble burst. The S&P 500 itself isn’t perfect at representing the market, but it is vastly better than the Dow.

The result is that the S&P and the Dow move in opposite directions more often than ever before. Since the S&P was created in 1957, one rose while the other fell on average one day in 10. Now they diverge one day in four—higher even than during the frantic trading of the dot-com bubble in 1999-2000 or the bond-market massacre of 1994.

If there is a crash in the expensive AI stocks, investors who followed the Dow would be far better insulated than those in the S&P. But those who want to diversify away from the Mag 7 have a much better alternative in buying an ETF that tracks the equal-weight S&P.

This is equivalent to buying the average of the top 500 companies because the index puts the same weight on Apple, with its market value of $3.4 trillion, as on the $7 billion car-parts maker, BorgWarner—and, as luck would have it, at the moment is similar to the Dow.

Write to James Mackintosh at james.mackintosh@wsj.com