as long as they change differentially. Thus, in 1938 stockholders at the top of the income scale were more heavily "overtaxed" than in 1951, despite a corporate rate of 50.75 percent in this latter year compared with 1938's 19 percent, and a rise in personal marginal rates from 74 in 1938 to 91 in 1951. For which Ce rose by 31.75 points, 1-P fell from 26 to 9, i.e., proportionately more than the increase in the corporate rate.

As already noted, the combination of a progressive personal tax and a flat corporate tax yields an "extra" burden of decreasing severity as stockholder income rises. Now we can, from this brief historical review, add this additional information: that, over the last 20 years or so, the "extra" burden on distributed earnings grew severely for most income-tax payers, but rose sluggishly, if at all, for those at the top of the income scale.

Now back to our main interest-the role of the relief provisions of the Internal Revenue Code of 1954.

The effect of the dividend exclusion and tax credit

The tax burden on dividend recipients was lowered in two ways in 1954. President Eisenhower originally proposed relief of this order of magnitude:

Specifically, I recommend that the credit be allowed on an increasing scale over the next 3 years. For this year, I recommend that a credit of 5 percent be allowed; for 1955, a credit of 10 percent; and, in 1956 and later years, 15 percent. To avoid shifts in the payment dates of corporation dividends, these credits should apply to dividends received after July 31, of each year. To give the full benefit immediately to small stockholders, I recommend that the first $50 of dividends be completely exempted from tax in 1954 and that the first $100 be exempted in 1955 and later years.*

This proposal proved to be one of the thorniest and most controversial considered in writing the revenue bill. After hearings and debate, Congress followed the outlines of the President's suggestion but set the amounts at a lower level. Marion B. Folsom, then Under Secretary of the Treasury, noted:

Under the new code each stockholder will be permitted to exclude from his gross income up to $50 of dividends and will be allowed a credit against tax equal to 4 percent of the dividends in excess of the exclusion. The amount of the credit is limited to 2 percent of the stockholder's total taxable income in 1954 and to 4 percent in later years.5

N

E

Remember the formula that summarized the "extra" burden, viz,

=

C. (1 - P). This now must be adjusted because of the relief provisions. First, we take up the credit. Now D E - CE. Therefore, the credit which is equal to 0.04D, also equals as a rate (i.e., after division by E) 0.04 0.04 (C.), and with C. equal to 52 percent, this comes to 1.92. So at all income levels, the "extra" bur

3 Or more precisely, to revert to our earlier symbols, as long as the change in C. and the change in 1-P are proportionately different. When C. increases and P falls, the differential will always increase; when C. falls and P rises, the differential will always decline. "The Budget of the United States Government for the Fiscal Year Ending June 30, 1955," Bureau of the Budget, 1954, p. M 18.

5 Remarks by Marion B. Folsom, Under Secretary of the Treasury, before the American Management Association, New York City, Aug. 19, 1954. The $50 exclusion applies to separate returns. Stockholders filing jointly are permitted an exclusion of $100, if each has at least $50 of dividends. See Internal Revenue Code of 1954, Public Law 591, ch. 736, secs. 34 and 116.

den is lowered by 1.92 percentage points. (It is worth noting that had President Eisenhower's original proposal been adopted, currently the credit would amount to 15 percent of dividends or 7.2 cents per dollar of earnings for distribution and this would provide a net tax benefit at higher income levels.) Because of the credit, we adjust downward our measure of the differential (the "extra" burden as a rate) against earnings for distribution; i.e., it now reads.

The credit, of course, will lower the differential by a flat amount (as noted above, under existing law, assuming the full corporate rate of 52 percent to apply, the differential will be cut by 1.92 percentage points). But the differential itself varies inversely with P, i.e., with the stockholder's taxable income. Thus, near the bottom of the income scale the credit would be responsible for a very slight relative reduction in the differential; near the top of the income scale, however, we should expect to find the "extra" burden reduced by a substantial percentage.

The credit, as we have just observed, gives a flat amount of relief at all stockholder income levels, and thus is not directly geared to the condition for which it is designed to provide relief. But the exclusion is even less focused on the problem. For, like any deduction, it is the more valuable the higher the applicable marginal rates of tax, while the differential declines in severity the higher the stockholder's marginal rate of tax. But limited to a specific amount of dividends, the exclusion has a strong effect only on those who receive a small amount of dividends. In our tabular comparisons below which are conducted on the basis of marginal dollars, we neglect the exclusion completely (assuming the marginal dollar to come in above the exclusion).

In formal terms, and considering all of a stockholder's dividends rather than the marginal dollar thereof, we can combine the exclusion and credit this way. With the corporate tax at 52 percent, earnings for distribution would be slightly more than twice as great as dividends, and the exclusion limits would then be $104 (i.e., the pretax equivalent of $50) for separate and $208 (i.e., the pretax equivalent of $100) for joint returns. For brevity, only joint returns (the majority) will be considered. The maximum relief afforded by the exclusion varies from $20 to $91, or from 20 to 91 percent of the excluded amount. With the corporate tax at 52 percent and dividends equal to 48 percent of earnings for distribution (designated as E), the relief provided by the tax credit equals 0.04 (0.48 E-$100) for all stockholders. Hence the combined relief, i.e., the sum of the credit and exclusion covers a span from 0.02 E+$16 for stockholders in the 20-percent rate bracket, to 0.02 E+$87 for those subject to a marginal rate of 91 percent; or, measured as a differential relative to E, from 0.02+$16/E to 0.02+$87/E.

This is the same as saying that since, with a corporate tax of 52 percent, D=0.48 of E, the credit of 0.04D equals 0.0192 or 1.92 percentage points.

7 For example, not paying tax on $1 of dividends saves 20 cents in the 20-percent rate bracket and 90 cents in the 90-percent rate bracket. Yet at the former level, the differential is 40 percent (assuming, for simplicity, a corporate rate of 50 percent) and only 5 percent at the latter level.

When E is small, say $250 (i.e., when dividends are $120), the fractions $16/E and $87/E will be considerably larger than 0.02 and noticeably different from each other. The exclusion feature will outweigh the credit. When E is large, say $100,000, the two fractions and the differences between them become insignificant. The credit predominates; the relief is very close to 2 percent of earnings for distribution. We cannot, therefore, simply conclude that the patterns of relief described for the credit and exclusion separately will characterize their combination. The degree of relief will vary with the amount of earnings for distribution. This, of course, has reference only to comparisons of average effective rates and differentials based thereon, but not to those comparisons concerned with marginal dollars of earnings for distribution.

So much for the general picture. The specific magnitudes that apply currently are summarized in table 2 which lists the differential tax rates on an added dollar of earnings for distribution:

1. As it would have been without any credit (column 2).

2. As it stood with the 4 percent of dividends tax credit (column 3).

3. As it would have been with a 15-percent credit (column 4).8

TABLE 2.-"Extra" burden computed on marginal rates and relief provided by dividend tax credits of 4 percent and 15 percent, at selected levels of stockholders' taxable income, 1959 1

1 Computed on the basis of a corporate tax rate of 52 percent and personal marginal rates applicable to Joint returns in 1959.

2 As originally proposed by President Eisenhower, see the Budget of the U.S. Government for the Fiscal Year Ending June 30, 1955, Bureau of the Budget, 1954, p. 1718.

"Extra" burden converted to a tax saving.

Because P is always positive, in the absence of any credit a positive (albeit declining) differential would exist at all income levels. The relief presently afforded by the credit is not sufficient to change this, involving at all income levels a cut of slightly under 2 points in the differential. Had the 15 percent tax credit been effective, however, the relief would have been enough to cause a negative "extra" burden on earnings for distribution at the top two of the taxable incomes

There are two reasons for this latter calculation. In the first place, some of its proponents had a higher level of credit in mind for the future when they referred to it as a modest step in the right direction. And, secondly, it is by a calculation of this sort that we can point up a difficulty inherent in the tax-credit device.

9

listed in our table. Any higher credit would mean a crossover at a lower level.10 A more liberal dividend credit would achieve neutral (equal) taxation of earnings for distribution at only one income level, leaving all others below this "overtaxed," and all stockholders above this point undertaxed.

The last three columns of table 2 are designed to illustrate a feature of the credit already noted. With a flat amount of credit per dollar of earnings for distribution 11 and an "extra" burden that falls with rising stockholder income, the degree of relief, i.e., the percentage by which the "extra" burden is cut by the credit, increases with stockholder income. Thus, the credit currently ameliorates less than 5 percent of the "extra" burden at the lower income levels (per marginal dollar of earnings for distribution), but relieves those at the top of the income range of over 40 percent of their differentially heavier taxload. If the 15 percent credit had been enacted, then, with tax rates assumed unchanged, at the lower stockholder incomes less than 20 percent of the "extra" burden would have been removed, while substantial relief would have occurred higher up-well over half at $100,000, and relief so great as to result in a tax saving for the very highest incomes.12

were to jump from the 15 percent credit way over Canada's 20 percent and on to considering what would be necessary in the way of a credit to remove the "discrimination" against the lowliest dividend-receiving taxpayer. (Even this, Utopian as it sounds, would leave nontaxable persons and institutions with an "extra" burden.) For to accomplish this something like an 87 percent credit would be needed. 13 With dividends eligible for credit running on the order of $9 billion, the revenue cost of this remedial measure to the Treasuy would be $7.8 billion. And despite this heavy cost, its accomplishments would still be highly qualified. Equitable tax treatment would not have been provided for nontaxable persons and institutions, and enormous tax savings would have been provided for those with taxable income subject to marginal rates of more than 20 percent." In contrast, simply exempting dividends paid from corporate tax, i.e., providing a dividend paid deduction at the corporate level similar to the current deduction permitted

More precisely with a 52 percent corporate rate assumed, the "crossover" from extra "burden" to tax "benefit" would have occurred at $180,000 of taxable income. This is for joint returns. For separate returns, subject to higher marginal rates and, hence, lower differentials, the "break-even" point would have come at $90,000 of taxable income. (For a neat and precise formulation of the relationships involved here see Carl S. Shoup, "The Dividend Exclusion and Credit in the Revenue Code of 1954," National Tax Journal, March 1955, p. 147.) In terms of our symbols we must find a P so that C. (1-P)=0.15 (1-C.). With C. at 0.52, the relevant P is about 0.86, the closest bracket rate to which is 0.87, applicable at the taxable incomes cited.

10 Canada's dividend tax credit, instituted in 1949 at 10 percent, currently is 20 percent. Assuming a corporate rate of 52 percent, the credit is 0.04 ($0.48) or $0.0192 per dollar of earnings for distribution.

12 Remember that this applies to marginal increments to given sized incomes, and the particular figures just cited also refer to joint returns only. For separate returns which were not, of course, permitted to split their income, substantial relief would occur considerably lower down the income scale as would the transition from "over" to "under" taxation.

13 Explanation: With the corporate rate at 52 percent, every $100 of pretax distributed earnings pays a corporate tax of $52 and, in addition, when the $48 that remains is distributed, an additional tax of $9.6 is forthcoming from a taxpayer in the 20 percent bracket. The tax in all, then, comes to $61.60; whereas, had there been no corporate tax and the full $100 had come out to him, his tax liability would have been $20. Thus, he bears an extra tax of $41.60. If he were given a credit of 86.7 percent of his dividend receipts, this extra tax would be completely alleviated-for 0.867 (48)=$41.6.

14 A marginal $100 of dividends at the top of the income scale (90 percent marginal rate) would pay a tax at this level of credit over $31 less than the tax due on an additional $100 of wages and salaries. The savings would dwindle, of course, as one reads down the marginal rate scale to a low of 1 percent of earnings for distribution for those subject to a 22 percent marginal rate.

for interest paid, would provide a completely equitable taxload on the distributed corporate earnings portion of their income for all stockholders taxable on nontaxable and would, in addition, cost the Treasury less-something like $6 to $6.5 billion at current dividend levels. (Using current dividend levels neglects the effect that such a change in tax law might have on corporate distribution policy and thus very likely understates the revenue cost.)

To stop at this point would leave the story seriously incomplete. All we have said so far relates to distributed corporate earnings only, but this is only a part of the broader problem of differential income taxation of stockholders. This is not the place for a careful development of the other considerations that are revelant here, but it would be inappropriate were we to say nothing at all about it. Therefore, we summarize briefly the considerations and findings relevant to a rounded treatment of the problem. The summary will be succinct and dogmatic. Anyone who is interested can find an expanded treatment of this range of problems and findings relevant thereto in the author's "The Income Tax Burden on Stockholders' published last year by Princeton University Press for the National Bureau of Economic Research.

The differential taxation of stockholders-A brief discussion of concepts and findings

1. The first step in rounding out our treatment of stockholder income taxation is to bring retained earnings into the picture. The "extra" burden on retained earnings can be measured by procedures similar to those used for distributed earnings. Specifically, it may be considered to be the difference between the potential personal income tax on the stockholder's pro rata share of earnings for retention (the amount of earnings prior to corporate tax that would enable a given amount of retained earnings net of tax) and the corporate tax actually levied on earnings for retention. The difference between these two tax liabilities is the "extra" burden if the actual exceeds the hypothetical; the benefit if potential liability is bigger than the actual. Following our earlier usage, the "extra" burden (or benefit) taken as a percent of earnings for retention is designated the differential against (or in favor of) earnings for retention.

Symbolically, it can be summarized this way. Add to the symbols used earlier:

Rearnings for retention

C, effective rate of corporate income tax on earnings for retention (this is higher than C. because earnings for retention are net of deficits reported by loss corporations).

N, absolute extra burden on earnings for retention

Nr

R

=

=

Then

differential against earnings for retention