Vance Archive   |   Trotskyist Writers Index  |   ETOL Main Page

T.N. Vance

The Myth of America’s Social Revolution

(May 1953)

From The New International, Vol. XIX No. 3, May–June 1953, pp. 167–172.
Transcribed by Ted Crawford.
Marked up by Einde O’Callaghan for the Encyclopaedia of Trotskyism On-Line (ETOL).

Shares of Upper Income Groups in Income and Savings
By Simon Kuznets, assisted by Elizabeth Jenks
National Bureau of Economic Research, Inc. 1953, 725 pp., $9.00.

The political economy the United States of America is indeed strange, as has frequently been remarked by analysts with varying points of view in the political spectrum. Moreover, in no other country is public relations and the art of sweeping exaggeration been carried to such refined lengths. This social environment helps to explain why a crude statistical work achieves front page publicity in the New York Times.

When the preliminary findings of the Kuznets study were released early in 1952, the New York Times gave them substantial coverage in its issue of March 5, 1952, starting with a front-page headline: Shift in Income Distribution Is Reducing Poverty in U.S. The lead paragraph by economic reporter Will Lissner stated: The United States has undergone a social revolution in the last four decades, and particularly since the late Thirties.” To be sure, the same newspaper, in an article by the same reporter one month later – to be precise on April 3, 1952 – carried an article with the headline: Living Standards Off 4 per cent Since Korea. This is the conclusion of a study by Dr. Julius Hirsh on the impact of price rises and tax increases on the moderate income city worker’s four person family – “the type of family ... that occurs most frequently in the varied structure of the American urban family.”

The “social” revolution apparently was not too profound, or at any rate it proved to be rather short-lived. Perhaps history was rather unkind to the advocates of the American “social” revolution by launching the Korean war before the findings of the Kuznets study were made public, and before the advertising agencies could use these findings to launch a campaign for reduction of taxes on the upper income groups.

What are the Kuznets’ findings? Lissner summarizes them with reasonable accuracy in the above-mentioned article, as follows:

As a result of little-appreciated changes in the distribution of a rapidly growing national income, the United States has gone about half the way toward eliminating inequities in incomes. But it has done this, not by leveling down, but by leveling up. These are some of the changes:

The very poor have become fewer by two-thirds of their 1939 number.

The poor have become better off. Where three out of four families had incomes of less than $2,000 a year in 1939, only one out of three fell into that class ten years later.

The well-to-do and the rich have become more numerous. In the late Thirties one family in about fifty was in the $5,000 and over income class, and one out of 100 was in the $10,000 and over class. In the late Forties, one family out of six was in the $5,000 and over class, and one out of twenty in the $10,000 and over class.

Over the years, the very rich have become poorer because the rise in labor incomes has been accompanied by a decline in property incomes. The share of the upper 1 per cent of income receivers in total income has declined in thirty-five years from 16 per cent to 9 per cent.

The Kuznets study, of course, is concerned primarily with what has happened to the upper income groups – the top one, five or seven per cent of the population. In his article in the May 1, 1953 issue of the New York Times, based on release of the entire study, Lissner provides a more up-to-date summary of the major findings of Kuznets’ statistical analysis and identifies the source of interpretation of these income changes as a “social” revolution.

The decline in upper group shares of total individual income was sharpest for the top 1 per cent of income receivers in the total population. This group had per capita incomes of $5,500 and up in 1948 and thereafter, and typical family incomes of $22,000 and up. Its share, before Federal income taxes, dropped from 12 per cent in 1939–40 to 8½ per cent in 1947–48. After taxes, the drop was from 11 to 6 per cent ...

From 1913 to 1948 the per capita income of the top 1 per cent little more than doubled. The Consumers Price Index rose two-and-a-half times its 1913 level; the upper group failed even to maintain its real income. The per capita income of the mass of the population, the lower 99 per cent group, rose to four times its 1913 level, making a vast improvement in its real income.

This was much more than a mere consequence of the shifts in income distribution which have been reducing poverty in the United States, reported in detail in The New York Times of March 5. These shifts, called “a social revolution” by Dr. Arthur F. Burns, Economic director to the President and research director, on leave, of the National Bureau would have produced only a moderate proportional decline.

Inasmuch as there have been more profound statistical studies than this, including several by Kuznets – none which has received notice outside the professional journals – one forced to the conclusion that it is label “social revolution” that is largely or exclusively responsible for widespread dissemination of the findings of the present study. And it is not without interest that Burns, also carries the title of Professor of Economics at Columbia University, is now chief economic adviser to the President.

Whether Burns is aware of the meaning of the phrase, “social revolution,” we do not know. Certainly Kuznets is not in any way responsible for this remarkable label. He merely presents his findings in a technical manner, hardly intended for the lay reader, surrounds them with the usual caveats and tables of derivatives and substantiation almost without end. The suspicion must remain however, that Burns was well aware of the fact that referring to changes in income distribution as a “social revolution” would result in extraordinary publicity and presumably in support for redistributing the tax burden – a goal that Burns apparently favors. Consider, for example the following paragraph from the first Lissner story:

He, Arthur F. Burns, who directed an important part of these investigations, concludes that we have about reached the limit of the usefulness of the income tax as a device for redistributing income. To raise the large revenues required for security at home and abroad, the tax must be heavily on the brackets where income in concentrated – moderate-sized incomes.

The “social” revolution thus fades into something far short of the expropriation, or even the impoverishment of the bourgeoisie. It would seem to center around the high individual income tax rates and the reduction in the proportion of national income going to dividends and interest – developments flowing from the development of the Permanent War Economy. The most important development of the Permanent War Economy, in so far as Kuznets’ findings are concerned, is clearly the sharp reduction in unemployment.

States Kuznets (p. xxxvii of his Introduction and Summary):

This recent decline in upper group shares, which for its magnitude and persistence is unmatched in the record, obviously has various causes. The most prominent are the reduction of unemployment and the marked increase in total income flowing to lower income groups (particularly farmers and wage earners); shifts in the saving and investment habits of upper income groups which may have curtailed their chances of getting large receipts from successful mature capital and equity investments; lower interest rates; and steeper income taxes. But conjectures alone are possible, and the discussion in the report is limited to a statement of facts.” (Italics mine – T.N.V.)

It is more than a coincidence that the basic economic program of the Eisenhower Administration is to reverse this so-called “social” revolution by reducing taxes on the upper income groups, raising the rate of interest, stimulating venture capital and thereby encouraging higher dividends, and stimulating a slight case of unemployment so that labor will not be so demanding and wages can be reduced.

Only the exigencies of the class struggle can account for the absolutely unpardonable use of the term “social revolution” in connection with the relatively insignificant changes that have taken place in income distribution since the development of the Welfare State and, more recently, the Permanent War Economy. Nevertheless, it is still of considerable interest to examine the changes that have taken place in the distribution of income.

Of more interest than the findings of Kuznets are the reports of the Census Bureau. These are based on Census surveys and may be considered to be much more reliable than data based on income tax returns, as is true of Kuznets. The Census data are before taxes and limited to wage or salary recipients. Dividing the latter into five groups, we get the following picture in percentages for selected years from 1939 to 1951:

Wage or Salary










Lowest fifth








Second fifth








Middle fifth








Fourth fifth








Highest fifth








In other words, so far as wages and salaries are concerned, accounting for about 70 per cent of total income payments to individuals, the middle income groups have gained – not only at the expense of the upper income group, but also at the expense of the lower income group. At any rate, regardless of what interpretation one cares to make of the above figures, there is clearly nothing that can justify the use of the term “social” revolution.

Kuznets, of course, is concerned primarily with the upper income groups. His figures show a higher decline for the top 1 per cent than for the top 5 per cent – and it is clear that no definition of the upper income groups can properly extend as far as the top 20 per cent. But the major decline has taken place since 1940–41, and this is precisely the period in which individual income tax rates have been raised enormously. The question of the reliability of the estimates is an inevitable one, and Kuznets is greatly bothered by it, spending an entire chapter of 75 pages, including appendix tables, in justifying his methodology. The chapter starts, however, by stating (p. 435):

We cannot measure the probable errors in our estimates directly because our basic data are either by-products of tax administration or products of censuses, subject to all the imperfections of social records. Some defects are obvious and the adjustments discussed in preceding chapters were designed to correct for them as far as possible. But after all these adjustments, errors inevitably remain, and we are faced with the difficult task of appraising them. This discussion of the reliability of our estimates must necessarily be incomplete and inconclusive. (Italics mine – T.N.V.)

If it is inconclusive as to whether the estimates are reliable, it may be wondered why the study was made. Kuznets indicates that the choice between using income tax returns and abandoning the study, and he obviously feels that the basic trends revealed by his study are correct. If by this were meant the small relative improvement in the position of the middle income groups, as shown by Census data, empirical evidence would clearly confirm such findings. For the average number of income earners has increased sharply among factory and white collar workers’ families as unemployment has decreased and the percentage of women employees has risen to an all-time high. In other words, on a family basis there can be little doubt that there has been increase in the average standard of living since 1939. This is also true on a per capita basis, but it is not so pronounced.

When, however, the claim is made that the upper income groups one per cent or five per cent) have experienced both an absolute and relative decline in their income shares, therefore presumably in their standards of living, one should look with a rather skeptical and jaundiced eye on an analysis that depends completely on the reliability of income tax data. After many comparisons and reliability tests, Kuznets refers to a sample audit study of 1948 income tax returns (which show a minimum of 70 out of 100 returns in the $25,000 and over bracket as containing errors) and concludes (p. 466):

The audit study, as far as the recent results go, warrants an inference that such underestimation is within a 5 cent margin for incomes at the to 5 per cent level, and within a 10 per, margin for incomes in the 2nd through to 5th percentage bands. (Italics mine – T.N.V.)

The difficulty is that the results do not go very far. They cannot do justice the extensive legal tax avoidance practiced by the upper income groups as analyzed in some detail in Part VI of The Permanent War Economy: Taxation and the Class Struggle, cf. November–December 1951 issue of The New International).

Our own private sample study of millionaires (the only reliable method of estimating what has happened to the incomes of the bourgeoisie) indicates that they are managing to survive although the fees to tax accountants and lawyers have increased rather sharply. Mansions costing in excess of $100,000 are still being built – in fact, in larger numbers than in any period during the last 25 years. Of course, vacations are frequently transformed into business trips – or is it vice-versa? Profits are frequently allowed to remain in corporations, in the expectation that the Eisenhower administration will ultimately reduce the surtax rates in the upper income brackets, so that it will “pay” to retrieve the dividends that are waiting to be declared. Some of these factors Kuznets tries to take into account, but the majority (and they are cumulatively decisive) are beyond statistical analysis.

We can only conclude that in a period of high tax rates any analysis of upper income groups based on tax returns is not only necessarily inconclusive, but tends to be unreliable. Kuznets, moreover, bases his analysis on a per capita approach. Aside from certain statistical difficulties in converting income tax returns to a per capita basis, the procedure as a measure of what has happened to upper income groups is exceedingly questionable. While the size of families in upper income brackets is smaller than in lower income groups, an upper income group with a large family might well be excluded from Kuznets’ array of the data on a per capita basis. If the purpose of the study is to discover something about standards of living, and not just to collect a lot of figures, then the facts of economic life have to be considered. Using the Kuznets approach, a single individual with an income of $25,000 annually would be part of the upper one per cent in 1948, but a family of five with the one income earner admitting to an income of $100,000 for the year might be excluded since the per capita is only $20,000. Such an analysis overlooks the fact that one mansion is usually sufficient for a family of this type; in any case, five mansions are rarely used. An analysis of shares of upper income groups necessarily involves a ratio of two quantities. The numerator, of course, consists of the amount of income going to the upper income groups, however income is defined. And it makes quite a difference as to what is or is not included in income. The Kuznets data necessarily contain a downward bias (probably on the order of twenty to thirty per cent) in the amount of income currently (since 1943) going to the upper income groups. The numerator of the income ratio is thus understated. But the ratio also depends on the size of the denominator. Here Kuznets uses what amounts to his own estimates of national income. This tends to overstate because of its inclusion of income in kind, imputed rent and other such concepts that are clearly not part of any analysis of the performance of a capitalist economy. If the numerator is noticeably smaller than it should be, and denominator somewhat larger than is proper for analysis, the resulting ratio is necessarily considerably smaller than it ought to be.

Unfortunately, we do not have available the statistical resources of the National Bureau of Economic Research or the Department of Commerce, but the decline in the shares of upper income groups since 1939 is not nearly as large as reported. Such decline as has occurred, moreover, is principally confined to the period upper income groups since 1929 is not Permanent War Economy. [This is in the original. – Note by transcriber] The bourgeoisie have not been destroyed or impoverished. They have succeeded, so far, in preserving their basic wealth, income and property. Nor has there been any diminution in the political power of the American bourgeoisie. What has happened, as we pointed out in the November#8211;December 1951 issue of The New International (p. 338), is that: “The state however, whose function is more and more to protect the rule and the wealth of the bourgeoisie, is being financed in steadily increasing measure by the workers and lower middle classes. Therein lies the secret role of taxation under the Permanent War Economy, while equality of incomes remains just as much a mirage on the horizon as it ever was.” (Italics in the original.)

Kuznets has contributed data that may be useful to income analysts. He is the real pioneer in national income data, and as one who justifiably claims to be a scientist in his field, he should blush at the “social” revolution that Burns has produced from his highly qualified data. Above all Kuznets ought to investigate why his data are being used as part of the drive, spearheaded by the NAM to reduce the tax burden of the upper income groups.

T.N. Vance

Vance Archive   |   Trotskyist Writers Index  |   ETOL Main Page

Last updated: 27 February 2019