If the White House has its way, investors will no longer have to pay taxes on dividends. The centerpiece of its much-anticipated $670 billion economic stimulus package is the abolition of the taxation of dividends at the investor level. How will this proposal affect investors and, specifically, investors in REITs? To give you a preview of the next 4,000-or-so words, we think this is a bear-market provision that will put a short- term halo on companies that pay dividends, with attendant boosts to their stock price, but that the market’s continuing addiction to (hoped for) capital gains will render any boost short-lived. Moreover, any boost would be susceptible to other, more pressing market realities. We believe it is unlikely that corporations will significantly change their policies toward capital formation or capital allocation as a result of this change. And we anticipate that regardless of the final form of the dividend tax and its treatment of dividends from real estate investment trusts, well-run mortgage REITs will still offer compelling total returns compared to other corporations.

Not to be overly simplistic, but investors have choices. For example, one can invest in stocks, bonds, commodities, real estate, art, cash, or anything that can gain or at least hold value, provide income, or both. There are sectors within these choices—growth or income, municipal or corporate, precious metals or grains, commercial or residential, Impressionists or photography, money market or checking deposits. And then, there are picks within the sectors within the choices—technology or biotech, utilities or preferreds, Monet or Lichtenstein. The asset allocation decisions made by investors are driven by any number of factors, from the quantitative to the ridiculous. But for stocks, over the past decade, growth has become the more highly-valued investment criterion over income. Chart 1 shows the average dividend yield on the S&P 500 has plummeted to below 2%:

It hasn’t always been this way. In fact, except for the last decade or so, dividends have done all the heavy lifting in total return to the market. Here are a few different ways to look at it:

• According to Bloomberg, from 1935 to 2001, $1 invested in the S&P 500 would have grown to $84.48 in simple price appreciation, and $1,180.79 with dividends reinvested. Without dividends, the average annual return comes out to 7%, with dividends it is 11.3%.


• According to Jason Trennert and Jim Moltz of ISI Group, the contribution of dividends to the market’s total return has averaged 51% since 1940, but in the 1990s only 23.2%. In the second half of the 1990s, capital gains dwarfed dividends even more. Chart 2 tells the story:
  
  
  
  
  
• According to the excellent research of Robert D. Arnott and Ronald J. Ryan of First Quadrant, in 1925, investors paid 18 years’ worth of then-current dividends to buy stocks; currently, investor’s are paying 80 years’ worth of current dividends to buy stocks, more than 4 times the levels of 1925. In other words, a dollar of dividend is worth a quarter of what it once was.


• Going even further back, according to Arnott, the average dividend yield on stocks since 1802 has been 4.9%. As recently as 1984 the dividend yield was about 5%. Since then, investors have put less and less value on the amount of dividends paid.


• Arnott’s research shows that $100 invested in 1802 would have grown to $700 million over the next 200 years, with dividends reinvested and after inflation. Back out inflation, and the $100 grows to $33 million. Back out dividends, and the $100 would have grown to just $1,884. To hammer home the point even more, the bulk of the growth from capital appreciation occurred in the bull market that began in 1982.


• Of the 8.0% annual total return of equities during the 200 years from 1802-2002, says Arnott, 5% is from dividends, 1.4% is from inflation, 0.6% is from falling yields and rising valuation levels, and 0.8% is from real growth in dividends. "It is self-evident," writes Arnott, "that dividends not only dwarf the importance of inflation, growth and changing valuation levels, but that dividends dwarf the combined importance of inflation, growth and changing valuation levels."

Indeed, the current generation of investors has ignored the long sweep of historical data come to consider income to be next to irrelevant in buying stocks. Investors buying stocks would look at a no-dividend company much like it was a zero- coupon bond with a face value that would only increase over time (how wrong they were). The lack of attention paid to dividends—or, more accurately, the declining dividend yield—is attributable to the incredible (and, in retrospect, bubble-type) capital gains that were racked up in the 1990s. (Remember, dividend yields are calculated by dividing the annualized dividend by share price, so as the share price denominator gets higher, the dividend yield gets lower.) Despite three consecutive down years in the stock market, valuation by virtually every objective measure remains high.

But companies have choices, too, when it comes to utilizing their capital—they can invest in the business, buy back shares, retire debt or pay out dividends. Companies have gotten the message: Dividend payout ratios, currently 32.8% for the S&P 500 after averaging over 50% in the past, are at historical lows. In 1977, more than 72% of the 7,000 companies tracked by S&P paid dividends, and in 2001 only 39% did. Now, 140 companies in the S&P 500 pay no dividends. Chart 3 presents the payout ratio by decade.

But times are changing. Capital gains are hard to find (the favored euphemism is "it’s a stock picker’s market") and income, specifically dividends, are regaining their lost respect. Now the White House wants to ride dividends’ coattails. President Bush announced today a sweeping plan intended to stimulate investment and the economy. Central to the plan is a call to eliminate taxes on dividends at the individual level. According to the White House, "The IRS taxes a company on its profits, then it taxes the investors who receive the profits as dividends. The result is that for every dollar of profit company could pay out in dividends, as little as 40 cents can actually reach shareholders…The President’s plan would eliminate the double taxation of dividends for millions of stockholders—allowing taxpayers to exclude dividend payments from their taxable income—and returning about $20 billion this year to the economy."

The New York Times described the Administration’s objective in changing this particular tax policy: "The proposal…is intended to stimulate the economy and reduce what many economists say is an incentive in the current law for companies to avoid paying dividends and run up debt." Reuters said that the proposal was "aimed at shoring up the economy and stock market." Given the information contained in Charts 1 to 3, which sets forth the emphasis—or lack thereof—placed on dividends by individuals and corporations, we have to ask whether the proposed legislation will have any of the desired effects. In other words, will cutting the personal tax rate on dividends from about 40% (the highest marginal tax rate on ordinary income at the federal level is 38.6%) to zero change the behavior of individuals and corporations such that the economy is stimulated and the stock market is boosted?

Changing Investor Behavior

We think that change, if any, will be ragged, short-term and incremental. First let’s discuss the first order effect of this change, i.e., a change in the behavior of the recipients of dividends, investors. There are two changes that we believe would be desired by the Administration. First, investors will understand that the pre-tax dividend yield is now also the after-tax dividend yield and, in turn, increase their demand for those stocks that pay dividends. Second, investors will not only keep—and spend—more disposable income from their dividend checks, but they will also see the increased capital gains from their existing holdings as the market re-values dividends.

All of this could happen, but we think the extent of the effect is debatable. Right off the bat, we would suggest that with a very weak employment and economic backdrop, it is likely that any extra disposable income from a dividend tax break might be saved or used to pay the gas bill or pay off credit card debt rather than used for consumption. Furthermore, increased taxes from budget-strapped states and cities would certainly offset any of the positive effects of this tax change. And if the tax change doesn’t go into effect retroactive to dividends received in 2002, then there will be no immediate increase in cash to the investor until 2004 as there is no withholding tax on dividends.

From a market perspective, over the last few years any investor who has been long stocks through a diverse portfolio of holdings or index funds has been seeing a lot of negative signs in his or her portfolio. A backward-looking recast of past returns to take into account the different tax structure would not, we believe, change anyone’s behavior. After all, in 2002, the S&P 500 lost 21.4% in simple price change, but after reinvestment of dividends the total return barely budged, to minus 20.2%. Eliminating taxes on the dividend component of the total return would not change anyone’s behavior, we think. The better way to look at the change is to look ahead to 2003 and see how returns would be affected using certain return assumptions, and then tax-effecting them. Arnott and Ryan point out that stocks have returned about 8.4% annually from 1926-2000. Let’s round up and say, for example (thank you, Jason Trennert, for inspiring this example), that over the next year the expected dividend yield is about what it is now, or 2%, which would mean that earnings for stocks (and therefore price appreciation) would be about 7%, for a pre-tax total return of 9% on stocks. Using the current 20% tax rate on long- term capital gains and a 40% tax rate on dividends, the after-tax return on stocks would be 6.8% (7% x 80% + 2% x 60% = 6.8%). Eliminating the tax on dividends would boost the after-tax return by 11.8% to 7.6%. In addition to the higher returns, if we put real numbers to this example, it also means that an investor with a $100,000 portfolio of stocks would have an extra $800 in his or her pocket.

Secondly, who would be benefiting from the tax change? The bulk of the benefits will go to the nation’s wealthiest investors—it has been estimated that 42% of the tax savings will go to the wealthiest 1% of taxpayers, and that 75% of the benefits would go to people with incomes of $100,000 or more—people for whom that extra dollop of spendable income has little resonance, even if it is used for consumption. Further, about half of dividend recipients are tax-exempt entities to begin with—foundations, pension funds or 401(k) plans—so the new tax break wouldn’t apply to them. So regulation won’t be applied efficiently. These investors won’t be rewarding dividend-payers just because they are marginally more tax-efficient.

Thirdly, it is possible that the tax change will flow through to stock prices as the market revalues, at most, the "new" dividend to the "old" after-tax dividend yield levels. The Reuters comment that the White House is hoping that slashing dividend taxes will help boost stocks is worth pondering further. We have read that the new dividend tax policy should increase stock prices by 8% to 10%. The stock market currently values dividends at about 2% pre-tax and 1.2% after-tax, so if the after-tax dividend has just been changed to 2%, will the market reprice upwards completely, thereby making dividend paying stocks more expensive? For example, let’s say Company X is a $10 stock that pays out 20 cents a year in dividends for a pre-tax dividend yield of 2%. An investor will keep 12 cents after tax under the current tax structure. Under the new tax structure, an investor will keep all 20 cents; however, if the market moves up because of the new dividend tax structure to the exact point where the after-tax dividend yield is equivalent, the value of the stock should be $16.67. In this example, without a change in the amount of dividend paid out the higher price brings the new dividend yield on this stock down to 1.2%, or exactly what the after-tax yield was before the tax was abolished.

Will the change to the taxation of dividends bring about a change in investor behavior? In an efficient market, yes, in the short run. But we do not believe that markets are efficient, that regulation is efficiently applied, or that investors behave rationally. Although the news about the reshaped dividend tax will be met with great gusto by the press and touted by the Administration, we’re not convinced that investors will change their behavior considerably or for very long. The change would require investors to start thinking about dividends on an after-tax basis. Dividend yields are currently quoted and discussed on a pre-tax basis. The municipal bond industry has had tremendous success over the years promoting the concept of taxable-equivalent returns in comparing munis to corporates or Treasurys, but we find it difficult to imagine a stock broker saying to a client, "Sure it’s just a 2% dividend yield, but now you get to keep all of it!"

In other words, in an efficient market the price of shares in Company X would see a one-time jump as the tax effect is factored into current valuations, all other things being equal—the extent of that jump is debatable; clearly other valuation benchmarks will keep the price down. We suspect that this effect may be mathematically inexact and, depending on the overall market, shortlived at that. We are reminded of a comment we read recently in Grant’s Interest Rate Observer: "[B]ooms have consequences that cannot be eliminated by central planners pulling a few levers."

Changing Corporate Behavior

This last point brings us to the other player in the dividend game—the corporation paying the dividend. Will the change in the dividend tax structure change the way capital is deployed? Our answer is again equivocal. It may provide incentive for those companies not paying dividends to begin paying them for short-term boosts. For example, a company like Microsoft, which pays no dividend, would get no boost from investors, and may choose to change its dividend policy if Bill Gates believes that investors are rotating out of MSFT into a company that pays a dividend. For companies that are already paying out dividends, a company that is paying a higher dividend yield should not result in a greater bump in valuation than one paying a lower dividend yield if the end result is a revaluation to the old after-tax dividend yield valuation. For example, as we saw above, assuming a complete discounting of the new tax to the new valuation, Company X, a $10 stock with a 2% dividend yield, would be revalued upwards by 67%. (Again, this is a mathematical exercise; such an increase would be bound by other valuation benchmarks.) Company Y, a $10 stock with a 3.5% dividend yield, would still be revalued upwards by the same 67%. See the table below:

Let’s say that Company X determines that there has been a favorable shift in market perception to dividend paying stocks and decides to increase its payout to 30 cents per share. Perhaps current holders would be happy because they get an extra ten cents in dividend. Perhaps the stock would get a one-time pop from a discounting of the higher dividend. Other than a possible bump to the stock, there is no real incentive for a company to initiate or increase its dividend. On the other hand, the market may punish Company X if it is thought to be a growth company. Consider the fact that Cisco shareholders last month rejected a proposal that would have required the company to pay a dividend for the first time. CEO John Chambers said after the November 19 vote, "Many people want a dividend because they’re worried about the growth rate of the company." Indeed, the problem facing the Microsofts, Ciscos and Suns of the world is that initiating a dividend may be perceived as abandoning growth. An editorialist in the Wall Street Journal put it succinctly when he said, "I avoid companies that pay dividends like the plague…[b]ecause when they pay a dividend they are admitting they have nothing better to do with their money."

This particular point of view may undergo a transition, but we doubt corporate managers will disagree with it. This is because another pernicious effect of a system that promotes higher dividend payouts is that it effectively reduces the funds available for the other uses of company cash flow—share buybacks, debt retirement or investments. The use of corporate capital is a zero-sum game, and to the extent that dividends are encouraged, capital spending is discouraged. (A cynic might even suggest that the change in the taxation of dividends gives room for corporation to cut the dividend.) It remains for econometricians to model the extent to which the perceived economic benefits of the new dividend tax structure would be offset by a reduction in business investment.

One idea that was discussed but left out of the final form of the stimulus package was to address the double taxation of dividends not only at the individual level, but also at the corporate level. In this scenario, corporations would receive a deduction for dividends paid, not unlike the way interest payments on debt are an above-the-line expense. This kind of proposal would have been much more likely to change corporate behavior and result in increased dividend payouts (CFOs are always on the lookout for ways to pay less in taxes), but we don’t think it was a particularly tenable or politically palatable idea. Most importantly, depending on the "success" of such a proposal, it would have a much more damaging effect on tax receipts than just focusing on the individual level. The Wall Street Journal estimates that deducting the tax on dividends at the corporate level would cost the Treasury from $40 to $60 billion per year, or approximately more than twice as much as halving the dividend tax for investors. We can’t imagine that the US Treasury would have been happy if a company like Microsoft, which has provisioned for a cumulative $20.7 billion in taxes over its last ten fiscal years, reduced its taxes paid, particularly in a year in which tax receipts are already falling so precipitously.

There will forever be academic discussions about the optimal capital structure for corporations. Some have argued that the tax break given to interest expense encourages debt issuance by corporations at the expense of dividend payouts. On the one hand, this may subject the borrower to market and fiscal discipline (a good thing), on the other hand it may promote excessive leverage (very bad). These two uses of cash aren’t equivalent, of course, because debt is presumably taken on for positive net present value projects, while dividends paid above the line would just reduce income without any return. After all, eliminating the taxes on dividends received does not really affect the cost of capital for companies. We will be the first to agree that the interest expense tax break, in the wrong hands, results in poor investments, serial acquisitions or excessive management compensation. But we will be the first to argue that it is doubtful that incentivizing managers to pay out more in dividends will make good actors out of bad ones, or make dumb capital allocators into smart ones.

Stocks, bonds and REITs

Regulatory changes affect behavior and behavior affects markets. This discussion of possible behavioral and market responses to a change in the tax on dividends has been necessarily inexact, because although we can predict how people and markets should respond, things rarely proceed the way they should. Abolishing the tax on dividends received should make investors seek out that kind of income—dividend-paying stocks outperform, corporations increase their payout ratios, investors rotate out of less tax- efficient strategies, and the market adjusts—but as we have outlined, the extent of the effect depends on the response of individuals and corporations.

It appears that the White House’s new stimulus proposal will exclude dividends from investments that already receive tax advantages due to corporate structure, like partnerships, sub-chapter S corporations or real estate investment trusts. REITs already avoid double-taxation of dividends: A REIT may deduct the dividends paid to shareholders from its corporate tax bill so long as the company’s assets are primarily composed of real estate held for the long term. The company’s income must also be derived mainly from real estate, and the company is required to pay out at least 90 percent of its taxable income to shareholders. The main benefit of being a REIT is one level of taxation while the main limitation is the restriction on the company to retain earnings.

Since their inception, REITs have provided competitive investment performance. Over the long-term, REIT market performance has been roughly comparable to that of the Russell 2000 Index and has exceeded returns on fixed debt instruments or direct investments in real estate. In recent years, REITs have outperformed most other asset classes. Because REITs annually pay out almost all their taxable income, a significant component of total return reliably comes from dividends. If we go back to First Quadrant’s return data showing the importance of dividends to historical total returns, if REITs had been around for the past 200 years, they would have vastly outperformed regular stocks.

Going forward, we believe that even with the prospective changes in the tax treatment of dividends, REITs will provide competitive returns to stocks. If the dividends received from REITs are not carved out from the legislation, then investors should see REITs enjoy whatever valuation boost the rest of the market receives. If dividends received from REITs are carved out from the legislation and remain taxed at regular income rates, then REITs and mortgage REITs in particular will still offer higher dividend yields and generally better ROEs than non-REIT corporations.

To use Annaly Capital Management (NYSE: NLY) as an example, since inception its total return, dividend yield and, most importantly, return on equity have been the envy of any corporation. Because of the REIT laws, Annaly isn’t a true equity—it is more like a cross between a leveraged bond portfolio and a financial intermediary. Its returns are competitive with bond funds on the one hand and thrifts and other financial institutions on the other. Investors are able to access a leveraged return in a tax-efficient manner and, in the hands of the managers at Annaly, that return has generally been better than equities. The table above illustrates how the total returns of Annaly have been superior to the S&P 500 for each of the last four years, even the bull market year of 1999, with most of that return in the form of dividends.

If taxes on dividends from non-REITs are abolished, it may boost their after-tax dividend yield, but not to anywhere close to the after-tax dividend yield of Annaly—currently over 14% pre-tax and almost 9% after-tax under the current tax structure—and other well-run REITs. Over time, the magic of compounding dramatically favors reinvesting the after-tax dividend of a REIT like Annaly over reinvesting the after- tax dividend of an S&P stock. Moreover, it won’t do anything for the operating efficiency and overall management performance of those companies. In short, if the market is going to start rewarding companies that pay dividends, the most efficient vehicle for that purpose is still a REIT.

The last point we’d like to make in this essay is that the primary motivation behind the Administration’s favoring a tax cut on dividends is economic stimulus. History has shown that the yield curve steepens and stays steep as the federal government and the Federal Reserve uses all of its tools to cure recessions (see "A Case for a Steep Yield Curve", October 16, 2002). This is an operating environment which bodes well for mortgage REITs like Annaly.

January 7, 2003
By Jeremy Diamond, Executive Vice President
Annaly Capital Management/FIDAC