Fortunately, figures appearing in the United States Bureau of Internal Revenue's report entitled "Statistics of Income" for 1942, pages 236-240, show, in one place, the total incomes of persons in the respective income brackets and also the total taxes paid by persons in the same brackets in each of the years from 1916 to 1942, inclusive. The figures are all there, together. These data, when converted to percentages of the national income, furnish an adequate basis for an analysis of the facts. You will find these conversions on two pages of the pamphlet herewith presented. The original figures from Statistics of Income have merely been divided by the national income figures here entered in the first column. I shall now present to your committee the results of this analysis. This is the first time that the findings have been made public. I believe that you will find that the results have a very important bearing upon the problems now before your committee. In the just-mentioned volume of Statistics of Income, information is also given showing both the effective tax rate and the tax rate applying to the highest segment of the income of each class. The latter is the marginal rate. Doubtless, you are well aware that, by the term "effective tax rate," the Bureau of Internal Revenue refers to the rate obtained by dividing the total tax receipts secured from the given class by the total income of the same class. If you will now turn to charts beginning opposite page, you will find the adjusted income figures plotted as historigrams, and you will see that, on the same charts, are shown the variations both in the effective tax rates and in the highest tax rates applying to the respective income classes. The graphs representing the highest rates are not quite as precise as are those portraying the effective rates, because sometimes the rates have been changed in the middle of the year. But they come as close as we could get to indicating the highest tax rates applying during the year. The effective tax rates are, of course, much lower than the highest tax rates applying to the same bracket. This fact is illustrated for the year 1947, by the table appearing on the back of the study which you now have before you. The CHAIRMAN. Where is that? Mr. KING. It is on the back of the study. That table compares the effective tax rate with the highest tax rate for the bracket. Federal tax rates on 1947 family incomes-Rates applicable to a family composed of husband, wife, and 2 minor children (recorded to the nearest whole percent) Mr. KING. It shows, for example, that, in the case of a family of four-a typical American family consisting of father and mother and two children-having an income of $3,000, the marginal tax rate is 19 percent, while the effective tax rate is only 5 percent. Similarly, in the case of such a family having a $10,000 income, the marginal tax rate is 29 percent, but the effective tax rate is only 19 percent. If you will now glance at chart 1-A, you will immediately note that a high degree of inverse correlation exists between the tax rates and the percentage of the national income going to persons in the $1,000,000-and-over bracket. For example, you will note that when, from 1915 to 1918, the tax rates rose sharply the percentage of income going to individuals in this class declined precipitously. Between 1921 and 1925, tax rates were lowered sharply. Between 1921 and 1929, the percentage of the national income going to this class grew tremendously. After 1931, tax rates climbed vigorously and, after 1931, the percentage of the national income received by persons having incomes of $1,000,000 or more remained at a very low level. Does this relationship characterize merely the highest income bracket? If you will turn to chart 1-G, you will observe that it tells much the same story. Evidently, this inverse correlation between tax rates and income is far from being merely accidental. If you will look closely, you will find that there was, apparently, a slight lag between the time when the taxes were levied and the time when the influence of the tax rates upon the incomes of the persons affected reached its maximum. Charts 2-A and 2-G represent in different forms the quantitative relationship existing between effective income-tax rates and the incomes of individuals in the respective income classes. In order to allow for the lag mentioned above, the income figure for a specified year has, in each case, been compared with the average of the effective incometax rates including that year and the 2 years just preceding. We experimented, I might say, with a number of other combinations, and we took the one that gave the best results-that is, the one that yields the highest correlation. On all of these charts, the relationships for the respective years are indicated by circles plotted at the appropriate points. Such an arrangement is referred to by statisticians as a "scatter diagram." It is so named because the effects of extraneous factors ordinarily cause the items to scatter widely over the chart field. This study is no exception to the general rule. It takes but a glance at chart 2-A, to make it obvious that low tax rates and high percentages of the national income received by persons in the $1,000,000 and over bracket go together, and that low percentages of the national income are associated with high effective tax rates. To aid the observer in visualizing, quickly, the general tendency indicated by the data, a 10-item moving median has been fitted to the plotted points. The moving median is indicated by the heavy line. The moving median has been employed for this purpose because it is the simplest and most effective mathematical device yet discovered for bringing out such general tendencies. It is more reliable than the moving average because the latter is unduly affected by erratic items. It is, of course, possible to fit to the data various mathematical curves, but most of these do not give results which are as dependable as these yielded by moving medians. However, the moving median has one unfortunate weakness: It tends to jump from level to level, and this tendency usually makes the graphs derived more or less irregular. The irregularity can be reduced by increasing the number of items in the moving-median group, but every such increase shortens the length of the line. The thing which you should now keep in mind is that the small fluctuations in the moving-median lines, as they appear in charts 2-A and 2-G, are due merely to extraneous factors, and have no significance as regards the analysis with which we are dealing. Observation indicates that in each of the charts 2-A and 2-G the moving-median lines show a marked tendency to slope sharply downward to the right. This means that, in each of these upper-income brackets, raising tax rates has tended to reduce the percentage of the national income going to the persons in the given bracket, while lowering tax rates has increased the percentage of the national income received by them. Just what is the explanation of this marked inverse correlation between effective tax rates and the percentage of the national income going to these upper brackets? To this question it is not easy to give a simple and categorical answer. Certainly the correlation is far too close to be accidental. Presumably, it is due to the fact that crushing tax rates lead people to work less diligently, and also prevent them from risking their capital in industry. If by a venture a man stands a chance of gaining several hundred percent on his investment, he is willing to take the risk of losing what money he puts in. However, if he knows that the Government will absorb in income taxes most of any profits which he may make, but will insist upon his taking all losses which may eventuate, he will not be foolish enough to put his money into any risky project. He will instead conclude that it is safer to invest his funds in taxexempt securities. Whatever the reasons bringing about the results, charts 2-A and 2-G, inclusive, prove definitely that high effective tax rates do reduce sharply the percentages of the national income going to individuals in the upper brackets, thus hampering capital accumulation, and retarding the upward movements of the incomes of all classes of citizensextending from wage workers to industrial magnates. Since high tax rates tend to make for lower incomes, the next question which naturally arises is whether or not high tax rates produce, for the Government, revenue greater than is obtained by low tax rates. The answer to this question is given by charts 3–A, 3-D, and 3-G. In these charts, the tax receipts expressed as percentages of the national income have been plotted against the average of the effective tax rates for the given year and the 2 years immediately preceding. Twelveitem to 16-item moving medians have been applied to the ordinates representing tax receipts. We stepped the number of items in the moving median up enough to get rather a smooth line. We just kept on increasing the number until we got a relatively smooth line which showed the general tendency. While the lines connecting these moving medians are a bit irregular, their general trend, nevertheless, shows clearly the connection between effective tax rates and income-tax yields expressed as percentages of the national income. The facts revealed by these moving median lines are amazing. Chart 3-A shows that, in the $1,000,000 and over income class, a 60-percent effective tax rate actually tends to produce less revenue than that obtained by a 22-percent tax. In the $500,000 to $1,000,000 bracket, the moving median indicates that a 23-percent tax rate gives as high a yield as does a 56-percent rate. In the $300,000 to $500,000 income class, the results obtained by a 26-percent effective tax rate are approximately the same as those given by a 50-percent tax rate. In the $150,000 to $300,000 income class, a 26-percent effective tax rate produced as much revenue as did a 43-percent rate. Chart 3-E, indicates that in the $100,000 to $150,000 bracket, raising effective tax rates from 19 percent to 32 percent added practically nothing to the Government's receipts. Only when we go to chart 3-F, dealing with the $50,000 to $100,000 bracket, do we find a distinct tendency for higher effective tax rates to produce higher revenues for the Government. In this instance, the moving median indicates that a 21-percent tax rate yields, roughly, one-fourth more revenue than does an 11-percent tax rate. Similarly, chart 3-G, referring to individuals having incomes from $25,000 to $50,000, shows that higher effective tax rates do bring in additional revenue. Here the indications of the moving-median lines are that a 13-percent tax rate produces some 60 percent more revenue than that obtained by a 6-percent tax rate. Apparently, then, raising effective taxe rates up to 20 or 25 percent may be correctly looked upon as a revenue-increasing device, but it appears that, after one passes the 25-percent effective rate, resorting to more intense taxation does but little to increase the volume of receipts pouring into the Federal Treasury. In the light of the facts revealed by this study, the conclusion seems to be that the levying of confiscatory taxes on high incomes accomplishes practically nothing in the way of raising revenue, but, as you are all well aware, is a very effective device for retarding the economic progress of all classes of the population. It, therefore, appears that such confiscatory rates cannot be justified on a fiscal basis, and can only be defended on the communistic theory that equality is more desirable than prosperity, or on the ground that success is criminal, and that those who succeed ought, therefore, to be punished. Since I am confident that no member of your committee endorses either of these philosophies, I feel sure that, in the light of the evidence, you will all be inclined to eliminate, at the earliest feasible moment, these oppressive and futile tax rates. Our committee will, in the meantime, use its best endeavors to make the public familiar with the facts, in the hope that, by so doing, it can gain for you the support which you need for sound tax revision. I wish to thank you for your attention. The CHAIRMAN. Any questions, Senator Byrd? Senator GEORGE. No questions. The CHAIRMAN. Thank you very much for coming and giving us the benefit of your paper, Mr. King. Mr. KING. Thank you, Senator, and I greatly appreciate your courtesy. (The charts referred to are as follows:) CHART 1-A HOW INCOME-TAX RATES HAVE AFFECTED 1939 Percentage of the National Income Received |