[Adapted from comments made to the Financial Industry Regulatory Authority in Washington, D.C., on October 15, 2007.]
One of the great unexplained curiosities of the mutual fund industry is its unwillingness to call attention to the vital role of investment income in shaping the returns on equities. Theory tells us, and experience confirms, that dividend yields play a crucial role in shaping stock market returns. In fact, the dividend yield on stocks has accounted for almost one-half of their total long-term return.
Of the 9.6 percent nominal total return earned by stocks over the past century, fully 9.5 percent has been contributed by investment return-4.5 percent by dividend yields and 5 percent from earnings growth. (The remaining 0.1 percent resulted from an 80 percent increase in the price/earnings ratio, from 10 at the start of the century to 18 at the end, amortized over the long period. I describe changes in the p/e ratio as speculative return.)
When we take inflation into account, the importance of dividend income is magnified even further (Figure 1).
During the past century, the average rate of inflation was 3.3 percent per year, reducing the nominal 5 percent earnings growth rate to a real growth rate of just 1.7 percent. Thus, the inflation-adjusted return on stocks was not 9.6 percent, but 6.3 percent. In real terms, then, dividend income has accounted for almost 75 percent of the annual investment return on stocks.
But while dividend income has accounted for nearly 50 percent of the long-term nominal annual return on stocks and 75 percent of the real annual return, even these figures dramatically understate the cumulative role played by dividends.
Consider this: An investment of $10,000 in the S&P 500 Index at its 1926 inception (Figure 2) with all dividends reinvested would by the end of September 2007 have grown to approximately $33,100,000 (10.4 percent compounded). [Using the S&P 90 Stock Index before the 1957 debut of the S&P 500.] If dividends had not been reinvested, the value of that investment would have been just over $1,200,000 (6.1 percent compounded)-an amazing gap of $32 million.
Over the past 81 years, then, reinvested dividend income accounted for approximately 95 percent of the compound long-term return earned by the companies in the S&P 500. These stunning figures would seem to demand that mutual funds highlight the importance of dividend income. But in this era of ''total return,'' income is virtually ignored.
Why? Because dividend income plays a remarkably small role in equity fund returns. Today, in fact, the average domestic stock fund is offering a dividend yield of just 0.4 percent. Where did all the income go? It was slashed by fund expenses.
Figure 1
Source: Bogle Financial Markets Center
Figure 2
Source: Bogle Financial Markets Center
The expense ratio of domestic stock funds averages 1.4 percent, reducing the funds' gross dividend yield of 1.8 percent to 0.4 percent. Unsurprisingly, then, it appears that the average stock fund earns the stock market's present dividend yield of 1.8 percent and then consumes fully 80 percent of that yield in fees and expenses. It didn't need to be that way.
When I began my research on this industry in 1950 for my Princeton University thesis, an interesting fact came to my attention. The first mutual fund-Massachusetts Investors Trust, founded in 1924-calculated its expenses not on the basis of a percentage of assets, but as a percentage of its investment income. During its first 25 years, MIT charged investors the then-standard trustee fee of 5 percent of income.
Throughout that quarter-century, MIT was the nation's largest mutual fund, and its growth was substantial. By 1950, its assets had grown to $362 million. The dividend income on its investments grew commensurately, and the 5 percent charge against income was soon producing far too much money for the fund's trustees to accept. (Imagine that!) So they promptly reduced the annual fee to 2.9 percent of income, by capping the number of shares on which the fee would be levied at 6 million.
Since dividend yields were then relatively high (MIT's stocks were yielding about 5.5 percent), the net dividend yield received by MIT's shareholders was 5.3 percent. (For the record, measured against fund assets, MIT's expense ratio was 0.33 percent.) For reasons lost in history, few of the mutual funds organized in the years after MIT followed the pioneer's precedent. Instead they chose to set their management fees as a percentage of net assets rather than as a percentage of investment income. The typical annual charge was set at 0.5 percent of assets, typically scaled down to three-eighths of 1 percent on fund assets in excess of $100 million. Modest fee structures, then, for an industry then managing modest amounts of assets.
A 1950 snapshot of that tiny mutual fund industry (Figure 3) shows both management fees and total expenses at a reasonably low level, along with a recognition by fund managers that, as their funds grew large (then, ''large'' meant more than $100 million in assets!), fund investors were entitled to share in the substantial economies of scale that accompany asset growth (i.e., that it cost little more to manage $200 million in assets than it did to manage $100 million).
But a funny thing happened. Those old values seemed to vanish. Amazingly, despite the truly staggering growth in total fund assets, expenses have grown at an even faster rate, resulting in expense ratios that have actually increased.
For five of these six funds, more and more of that priceless component of investment return known as dividend income was consumed by costs (Figure 4), from 10 percent of income in 1950 to nearly 60 percent in 2006. Even as assets have increased nearly 60 times over, from $770 million to $42 billion, their expenses have increased even faster—more than 100 times over, from $3.4 million to $395 million. Result: Expense ratios have nearly doubled, from 0.57 percent to 1.0 percent. This evidence totally contradicts the consistent stand of the industry, articulated over and over again at the annual membership meetings of the Investment Company Institute, that ''the interests of mutual fund managers are directly aligned with the interests of mutual fund shareholders.''
It's just not so. And now, a dream. Suppose now that industry practice had followed MIT's early lead, pegging management fees to 5 percent of investment income rather than to fund assets. Further suppose that no economies of scale—none—were shared with fund shareholders, and that the 5 percent fee remained unchanged. On that basis, equity fund expenses last year would have totaled just $5.7 billion, compared with the actual total of $56 billion, a huge potential annual ''dividend'' of $50.3 billion to fund shareholders. Well, I can dream can't I? But in any event, it's high time that we require mutual funds to disclose to investors and prospective investors the amount of their dividend income that is consumed by costs, and its impact on the fund's long-term returns.
Figure 3
Source: Bogle Financial Markets Center
*Now respectively, Lord Abbott Affiliated, AllianceBernstein Growth & Income, Putnam Investors and Vanguard Wellington
Figure 4
Source: Bogle Financial Markets Center
*Now respectively, Lord Abbott Affiliated, AllianceBernstein Growth & Income, Putnam Investors and Vanguard Wellington