By

The Dow Jones Industrial Average started with just 12 stocks in 1886 and grew to its current 30-stock roster in two big jumps: in 1916 the average was expanded from 12 to 20 stocks; and, in 1928, the final move from 20 to 30 stocks was made. Although it is easy to find out what companies are currently on the various Dow Averages, the actual mechanics for computing the Averages are much less transparent.

When Dow created his first Average (now the Transportation Average), he simply divided the total of all the prices by eleven (the number of stocks on the Average) to get his result. Living in the days before frequent stock splits and stock dividends, Dow did not foresee events that would make the divisor change. As markets matured and companies began to try to keep the price of their shares within certain parameters, stock splits became a common tool to keep the price per share at a reasonable level. Until the implementation of the Dow Divisor, Dow Jones would account for any splits by multiplying the stock price by the number of shares into which each share was split. Unless they did this, a two-for-one split in one of the stocks that made up the Dow would have caused a significant drop in the Average that was not the result of any fundamental change. For example, if General Electric had split 2-for-1 at some point in the past, one would multiply the current price of General Electric by two before figuring out what the Dow had closed at for the day.

Under this policy, the Average simply reflected the results of what a simple buy and hold strategy would have attained -- effectively neutralizing the effects of stock splits. However, stocks that split their shares numerous times would come to have more and more power over the daily moves of the Dow, because their small daily movements would be multiplied to account for numerous splits. This distorted the Average by making it "split-weighted" (allowing companies who split their shares more frequently to have a disproportionate control over the Average). To correct this, the editors of Dow Jones developed the Dow Divisor -- a single number that took into account the splits for each individual stock.

The Dow was first computed using the Dow Divisor on September 10, 1928. Norman Fosback describes how the Dow Divisor works in "Stock Market Logic":

"Assume that the prices of the 30 stocks summed to $3,000. Then the 30-stock average price would be $3,000 divided by 30, or 100.00. If one of the thirty stocks was quoted at $200, but then split two-for-one so that its price fell to $100, the sum of prices would be only $2,900. To maintain the Average at its actual level of 100.00, the divisor of the Average is systematically altered. Since $2,900 divided by 100.00 equals 29, the new divisor is set at 29, down from 30, in effect preserving the average price at $100 ($2,900 divided by 29 still equals 100.00)."

Dow Jones currently adjusts the Dow Divisor if any event effects the Dow Jones Industrial Average by 5.0 or more points. Stock splits are not the only factor that affects the Dow Jones Industrial Average -- stock dividends actually have an effect as well. When a stock pays a dividend, the "specialist" who facilitates trades in the stock at the exchange deducts the amount of the dividend, rounded to the nearest 1/8th, from the price of the stock. For example, if General Electric was trading at $50 and was scheduled to pay a $1 quarterly dividend on June 1st, General Electric would trade "ex-dividend" on June 2nd at only $49 a share and shareholders of General Electric would have the stock ($49) and the dividend ($1), for a total of $50 worth of stock and dividends. Large, one-time cash or stock dividends and spin-offs of subsidiaries can move the Average a heck of a lot more than 5.0 points, so the Divisor is also adjusted to compensate for all of this.

This Divisor approach worked well for the first few decades but, in recent years, the Divisor has started to become very small. In 1986, the Dow Divisor fell below 1.0 for the first time -- effectively becoming the "Dow Multiplier" since, to divide by a fraction, you invert it and multiply (resulting in a larger number).

For example, if the Dow Divisor were at 0.333 (it has many more decimal places, actually), the prices of the stocks on the Average would be effectively tripled when the Dow is computed. This creates a situation where even a fractional movement in the price of the Dow Industrial stocks (say up or down by 1/2 a point) results in relatively large movements in the Dow Average reported (a 1/2-point increase in each stock would translate to a 45-point move in the DJIA). See DJIA Divisors on the Dow Jones Averages website for the current divisor and a history of recent changes.

As a result of the way the Dow is computed, the Average has effectively become a "price-weighted" measure of the market. The Divisor simply measures the sum of the day's changes -- there is nothing in it to account for the fact that a $1 change in a $15 stock is, percentage-wise, greater than a $1 change in a $100 stock. Because it measures all net changes in the same way, in effect there is a weighting system that tilts movements in the Averages toward the higher-priced stocks. A $1 move up or down in a $100 stock is only a 1% move, a commonplace event, whereas a $1 change in a $20 stock is a 5% move which is more rare. The Divisor does not correct for the relative magnitudes of changes between differently priced stocks so higher priced stocks influence its movement disproportionately. Although this might make one question whether the Index is truly representative of the New York Stock Exchange (the original intent) or the broader market (its de-facto contemporary function), it helps us better understand the day-to-day movements in the Dow.

**History of the Dow**