When the Dow Jones Industrial Average crossed the 20,000 mark in January 2017, the media went wild and the president celebrated with a tweet. When it ground out another 25% to hit 25,000 on Jan. 4, 2018 – less than a year later – the cycle repeated itself. When the index dropped 4.6% to 24,344 on Feb. 5, you would have thought the sun had enveloped the earth.

To be sure, the stock market shot up in 2017, and to be sure, the recent plunge was steep. But we shouldn't describe market milestones or calamities in terms of the Dow (DJI). It's just not a good index.

Indexing, 1890s Style

Perhaps we should give the Dow a break. There wasn't much in the way of precedent for designing index methodology when it was founded in 1896; it was only the second equity index, and its inventor, Charles Dow, had also invented the first. Nor was there anything like the processing power necessary to support a complex index. If Dow had put every mathematician east of the Mississippi to work in unison, he could not have begun to match the computing might we wield by fiddling with our phones.

So should we let Charles Dow's average go the way of the telegraph, heroin-based children's cough remedies and other little-missed aspects of the 19th century?

The argument that it's time to retire the Dow starts with its size, just 30 stocks. Once upon a time that was a reasonable cross-section of the market. But today there are around 7,000 listed companies in the U.S. And it's not even as though the Dow's components are the biggest. Apple Inc. (AAPL) is there, sure. It joined in March 2015, a seemingly arbitrary time, since it had already been the world's biggest company on-and-off for four years. The second- and fourth-biggest companies, Alphabet Inc. (GOOG, GOOGL) and Amazon.com Inc. (AMZN), are nowhere to be seen, but number three Microsoft Corp. (MSFT) is, arbitrarily, there.

According to S&P Dow Jones Indices, which owns the Dow, "a stock typically is added to the index only if the company has an excellent reputation, demonstrates sustained growth and is of interest to a large number of investors. Maintaining adequate sector representation within the indices is also a consideration in the selection process." The components are not, however, chosen according to "quantitative rules."

Choosing blue-chips based on subjective criteria is forgivable, even if choosing only 30 of them might not be. It's what the Dow does next that makes it an inadequate gauge of market performance. The index is price-weighted: for every $1 one of its stocks goes up, the Dow goes up 6.89 points (the Dow divisor, which is used to keep stock splits from throwing the index off even further, yields that ratio).


Schwab's Terri Kallsen on Dow 20k

In other words, stocks with higher prices affect the index more, whatever their actual value as measured by market capitalization. At close on Jan. 3, Goldman Sachs Group Inc. (GS) was a $96.2 billion company by market cap with a per-share price of $253.29. Apple was an $894.0 billion company with a per-share price of $172.23. Apple is over nine times more valuable than Goldman, but the Dow gives Apple around two-thirds the weight it gives Goldman.

The result is that the Dow sometimes moves in the opposite direction to the S&P 500, another S&P Dow Jones index, which is capitalization-weighted and has 500 (well, 505) components. On Dec. 1, 2016, for example, the Dow gained 0.36%, while the S&P 500 fell 0.35%. If you had any interest in actually knowing how the market did on that day, you would look at the S&P 500 and realize that it was modestly down. If you were watching or listening to a non-specialist news broadcast, you'd be liable to hear that the Dow was up 68 points – sounds good, whatever that means. In other words, conflating the Dow with "the stock market" is stretching it.

An even worse practice is citing stock market gains or losses in terms of the Dow's points. The 1,175-point drop on Feb. 5 was not the largest drop in the history of the index by any (sensible) measure. Sure it sounds a lot bigger than the 508-point drop on Black Monday (Oct. 19, 1987) – but that was a 22.6% plunge, while the seemingly steeper 1,175-point drop only amounted to a 4.6% loss. And besides, the S&P was down 4.1%. Go with that.

Why the discrepancy between the Dow and the S&P? The Goldman-Apple example gives a hint: Goldman Sachs gained 3.3% on Dec. 1, 2016. Apple fell 0.9%. The S&P takes that information and concludes, "Goldman Sachs added around $3 billion in value to the market; Apple subtracted around $6 billion from it." The Dow? "Goldman Sachs went up $7.34 per share, so let's say it added around 50 points to the market; Apple went down $1.03 per share, so it subtracted around 7 points." (Note that the divisor has changed since these figures were calculated.) 

If that seems logical, we've got a SPDR Dow Jones Industrial Average ETF (DIA) to sell you.

What Can You Do?

David Blitzer, managing director and chairman of the index committee at S&P Dow Jones Indices, told Investopedia in January 2017, "There's certainly a lot of times when I'm as surprised as anyone else with how much coverage the Dow gets." Confronted with the issues caused by its tiny number of components and price-weighted methodology, he directed us to his company's other famous product, the S&P 500: it "doesn't have all the issues you just described." He added that professional investors and market analysts prefer the S&P 500, "but the Dow is the one that's all over the newspapers." (The Wall Street Journal, which Charles Dow co-founded in 1889, blared "Dow Hits 22000, Powered by Apple" – which isn't even a multiple of 5,000 – on Aug. 3.)

Blitzer provided some insight into the process of choosing the index's components. Because the Dow has long been accompanied by sister indices tracking transportation and utilities, it does not include companies from those sectors. Until the 1980s, the "industrial" moniker was more or less accurate, but then it was decided that the index would include banks, insurers, restaurant chains and other non-industrial companies – but not utilities or transportation. Blitzer also revealed a fudge the selection committee uses to avoid some of the price-weighted methodology's worst pitfalls: the ratio of the highest stock price to the lowest should be less than 10 to 1 (which may explain why Alphabet and Berkshire Hathaway are excluded). 

Blitzer appears to accept the issues with the Dow. He said that the committee has discussed switching to a capitalization-weighted approach. But "you'd just use the S&P 500 at that point" and, he added, "we'd be throwing away a lot of history."

Therein lies the rub. Fixing the Dow's problems would make it tough to distinguish from the more sensible indices that are already out there. In the process it would cut off a metric that, however flawed, can be traced continuously for 120 years – not to mention throwing away an internationally recognized brand. "Why do we keep it?" Blitzer laughed as he repeated the snide reporter's question. "The recognition, the publicity, the long history."

"To be honest," he added, "this is not a hugely expensive index to maintain."

If it's too broke to fix, don't. On the other hand, it's probably not wise to pay too much attention to it either.