Dividends Do Matter
By Harry Seneker

Illustration by Vasily Kafanov

Back to The Future Once upon a time, dividends mattered quite a lot to equity investors ¾ with good reason. From 1929 to 2002 the Dow Jones Industrial Average, to use a traditional proxy for large blue chips, produced an average return of about 11.31% a year. Of that 11.31%, about 6.9 points were capital gains and about 4.3 points, which is to say 38% of the average annual total return, came from dividends. And this includes the 1990s, when capital gains were rampant, and boards of directors and equity investors alike were often downplaying or even ignoring dividend payouts. By the 1990s taxes on capital gains were significantly lower than taxes on dividends, which were largely taxed at ordinary income ( i.e., relatively high) rates.

Now dividends are mattering again to many investable companies and to investors. Taxes and the rise and fall of speculative frenzy seem to have played a large role in this turnabout in attitudes. First came the bursting of the bubble, and then the recent recession. Capital losses widely replaced capital gains.

In 2003 a new round of tax cuts put capital gains and most corporate dividends (REIT dividends being one major exception) on equal and more attractive footing ¾ the maximum rate on both was set at 15% through 2008 (after which the rates are scheduled to revert to the previous higher rates if Congress fails to act). Meanwhile, Federal Reserve efforts to stem the recession forced interest rates to historic lows, causing the fully taxableyields on CDs and corporate bonds to drop to extraordinarily low levels. So corporate dividends, with their partial tax protection, have come to look much more attractive to many classes of investors. Which leads to the question, how best for investors and money managers to measure, select, invest in and benchmark corporate dividend-oriented investments?

Innovation Needed

It may seem there is already an index for everything in the equity markets, but that proved not to be the case here. To be sure, there were a number of indexes suggested as full or partial proxies for equity dividend investments, but all came up short. For example:

LARGE BLUE CHIP INDEXES, such as the Dow Jones Industrial Average, are likely to include stable, well-financed companies with excellent dividend coverage, but they will necessarily overlook numerous dividend payers merely because they are not gigantic in size. Moreover, they are hard put to leave out such a company as Microsoft, a modern blue chip by anybody's standard, whose dividend will be less than exciting to almost any income-oriented investor.

VALUE INDEXES, if broad-based, may capture a lot of dividend stocks, but again will include many with small or zero dividends. Value criteria typically focus more on low ratios to book value or earnings than on dividends. Value compilations also have the drawback of including companies with uncertain financial or operating prospects that investors seeking stable or growing income may prefer to avoid.

UTILITIES INDEXES have the merit of measuring a class of equities traditionally popular among dividend-oriented investors, but narrow the dividend-paying universe down to a single industry, undesirably limiting investment options and exposing investors to industry risk. Moreover, some utility holding companies falling under such a rubric have less-than-desirable recent records as safe investments.

BROAD EQUITY MARKET INDEXES may include nearly every name a dividend investor might want, but perforce will also include nearly every name such an investor would not want.

REIT INDEXES measure a category of equities that are in a class by themselves and therefore require special treatment. While something dividend-oriented investors may consider, REITs have very different tax treatment from other equity classes and carry exposure to real estate industry risk.

What Do You Need in a Divident Income Index?

BROAD COVERAGE. Consistent with Dow Jones Indexes' other indexes, the Dow Jones Select Dividend Index methodology starts with the largest feasible pool of investable equity names. The Dow Jones U.S. Total Market Index provides that pool, since its objective is to represent approximately 95% of the investable U.S. universe. As of June 30, 2004 , the TMI contained 1,633 components.

DIVIDENDS, OBVIOUSLY. At the time of the Dow Jones Select Dividend Index annual component review in December 2003, the Dow Jones U.S. Total Market Index contained 847 stocks paying regular dividends.

HIGH YIELD. The Dow Jones Select Dividend Index methodology then screens components to include those with the highest safe yields. The 847 dividend payers were ranked from highest to lowest.

TAX CONSISTENCY. The Dow Jones Select Dividend Index methodology then screens components to include those with consistent tax treatment of dividends paid. REITs, whose dividends are taxable as ordinary income to their shareholders and which have long been indexed separately by Dow Jones Indexes, were excluded from the universe of eligible issues.

STABILITY AND GROWTH. The Dow Jones Select Dividend Index methodology then screens components for stable, well-financed companies with safe dividends, preferably dividends with prospects for growth over time. This factor required some thought. A new index, in the view of Dow Jones Indexes, needs to be based on transparent, objective rules, not on subjective judgments, such as are often found underlying analyses of companies' prospects. Since income-oriented investors tend to put stability and safety of payout first, those issues were addressed first.

1. COMPONENTS MUST HAVE AT LEAST A FIVE-YEAR RECORD TO DATE OF PAYING REGULAR DIVIDENDS WITHOUT REDUCING OR SKIPPING PAYMENTS. Although past performance is not necessarily indicative of future results, a stable past dividend record is widely accepted as a starting point when seeking a stable future record. At the time of the Dow Jones Select Dividend Index annual component review in December 2003, that limitation produced a universe of 704 stocks with a good history of sustained payout. In the case where this rule failed, the equity would automatically and immediately be removed from consideration. But safety required more than that, leading the development team to further experimentation. After extensive research, Dow Jones Indexes established a second rule:
2. COMPONENTS MUST ALSO HAVE A FIVE-YEAR RECORD TO DATE OF DIVIDEND PAYOUTS NO GREATER THAN 60% OF AVERAGE PER-SHARE EARNINGS.

It has long been commonplace for established companies to pay out up to half of net earnings, with due allowance for maintaining payments in occasional poor-performance years. Pragmatically, companies that held bad years down to 60% payouts proved to be safer bets for sustaining dividends. And this standard easily eliminated most weak or troubled companies altogether. During the most recent annual component review, this restriction brought the count down to 574 companies. There was little problem in selecting the 100 best-yielding equities from this universe: 50 for the index, and a second 50 to provide a pool of replacement stocks.

INFLATION PROTECTION. This is commonly the concern that is a close second in importance to stability and safety of payout for income-oriented investors. A traditional argument for equity investments has been the opportunity to match or exceed general inflation. In practice, backtesting of the Dow Jones Select Dividend Index has revealed that its performance has been consistent with this objective.

LIQUIDITY. The Dow Jones U.S. Total Market Index has rules in place assuring a universe of investable securities for those who want to replicate it or use it as a benchmark for comparison purposes. However, the index contained 1,633 stocks as of June 30, 2004 , meaning the average weight of an individual stock would be 0.06%, and the recent maximum weight was 2.1%, for General Electric. For the 50-stock Dow Jones Select Dividend Index, the average weight would be 2.00%, with a maximum weighting in the vicinity of 4.94%. Greater liquidity would be required.

Dow Jones Indexes selected a $10 million Select Dividends replicating "basket" as a working standard to develop a minimum acceptable degree of liquidity. Since a typical individual stock weighting would be 1.5% (quite a few individual components falling below the average), that would imply a $150,000 investment in the average stock in such a basket. The aim was to be able to make such an investment in one day. It is normally thought that a competent trading desk can buy or sell 10% of a stock's daily trading volume without unduly "moving" the market. That would suggest that a stock must have an average daily trading volume of $1.5 million, which was thus set as the liquidity standard to be included in this index, unless and until practical experience indicates a higher standard is needed.

Weighting Dividends

A key issue in index development is how to weight the components. As already seen, a market-cap weighting scheme fails to give an appropriate weight to a company that pays a substantial cash dividend. Equal-weighting of components can be a useful alternative in many cases.

But the need to create an index that reflects a dividend-oriented investment strategy forced Dow Jones Indexes to think like a dividend-oriented investor, and to come to a very different conclusion. A dividend-oriented investor is interested in cash income. Market capitalization as such is not a significant issue once market liquidity needs are met. And weighting a component yielding a relatively lower dividend equally with one yielding a higher dividend made no sense at all.

If Dow Jones Indexes was going to innovate in the area of dividend orientation, it was going to have to innovate in the area of component weighting as well. Obviously, it made sense to weight such an index by the dividend payout ratios of the components, but that approach had problems of its own. It was one thing to select stocks for this index by yield; that was essentially a one-time event, and would be usually a once-a-year event for the continuing index. But the simplistic approach of ranking simply by yield had a flaw: Yield is a function of price, which fluctuates daily for most equities, sometimes dramatically so (Yield = Dividend/Price). Their yields necessarily fluctuate also. This approach would require frequent, perhaps even daily, rebalancing of the resultant index. That would sharply reduce, if not eliminate altogether, the value of the index for those replicating it for investment purposes or basing exchange-traded products on it (because of transaction costs and fluctuations in return), and greatly complicate the process of benchmarking against it. We needed to look for another approach.

Back to thinking like a dividend-oriented investor. Such an investor looks for high but safe yield. But once invested, he or she ideally receives a steady stream of cash dividends that changes in amount over time only slowly or not at all, whatever the prices of the stocks may do. And this stream is what he or she is primarily investing for. What if one chose componentsby yield, but weighted them according to the income stream the investor wants rather than the yield?

Components were weighted by the dollar amount of the dividend each paid. Immediately, the index stabilized. Now it required only the normal annual rebalancing (except for singular events like mergers, bankruptcies or a significant dividend reduction) common to virtually all equity indexes. Adjusting for stock splits or other capitalization changes was just as routine for a dividend-oriented index as it was for other equity indexes.

The proof of the pudding is in the testing

Dow Jones Indexes backtested the performance of the result of the index methodology over 12years, the longest period available for the underlying Dow Jones U.S. Total Market Index, comparing this index with market-cap weighting of the same components (since weighting the index by dividend yield had already been rejected for other reasons). The results were as follows:

For nine years of the first ten, dollar-dividend weighting produced a noticeably better payout than market-cap weighting, by amounts up to 17.1%. The one exceptional year, 1993, found it lagging by only 0.47%. Only in the most recent two of the 12 years has market-cap noticeably outperformed by this measure. An examination of the component data showed why: Virtually all the difference was accounted for by a single component, Altria Corp. (the company formerly known as Philip Morris). During that period it not only paid the highest or among the highest cash dividends, but also had the highest or among the highest yields and market caps among all components. The result: Altria won a 20%+ weighting by market cap, but received onlya bit over 4% under dividend weighting. The fivefold higher weighting skewed the resulting dividend payout.

In the view of Dow Jones Indexes, this one exception was not a reason to revert to market-cap weighting. Moreover, a weighting of 20% or more in one component in an index is not acceptable in an index designed as a benchmark for a segment of the marketplace presumptively pursuing safety, for which a well-diversified portfolio is a major requirement. Better to accept a lower but still quite good yield and income stream and have the diversification demanded. This consideration was particularly important in this case because the component in question is Altria, whose signature product, cigarettes, is considered a major causative factor in heart disease, cancer and emphysema. Whatever the eventual outcome for Altria, moreover, the historical record suggests that a list of large, reliable dividend payers dominated by a single top-payingcomponent is an anomalous situation, one that should not be counted upon to persist over the long run.

Dow Jones Indexes discovered through backtesting of the Dow Jones Select Dividend Index that the growth of its dividend stream outperformed inflation, and that its price-based return did the same. Dow Jones Indexes discovered the Dow Jones Select Dividend Index also outperformed vis-à-vis other indexes of income-yielding investments.

It turns out that Dow Jones Indexes' new Dow Jones Select Dividend Index would be a very attractive alternative to many investors and money managers seeking to outperform the conventional range of equity indexes under most market conditions that have been observed in the past 12 years. Dow Jones Indexes had not expected its new Dow Jones Select Dividend Index to shed new light on that ancient debate about the merits of old-fashioned, old-school investing ¾ for and against the school that calls on one to invest in solid, stable companies paying good dividends and hold them for the long term. But innovating to break new ground can sometimes take one in unexpected directions, and even yield new insights. This findingwas, however, an additional piece of evidence supporting something Dow Jones Indexes was not surprised to find:

Dividends do matter.

*This document contains comparisons, assertions, and conclusions regarding the performance of the Dow Jones Select Dividend Index based on backtesting, i.e., calculations of how the index might have performed in the past if it had existed. The Dow Jones Select Dividend Index was not designed, and was not calculated, between 1991 and 2003. As described in more detail at the end of this document, backtested performance information is purely hypothetical and is provided in this document solely for informational purposes. Backtested performance does not represent actual performance, and should not be interpreted as an indication of actual performance .