Knowledge that Transforms

I APPRECIATE the opportunity of replying to Professor Guthmann's gracious note. In the hope that some sources of confusion in my paper may thus be eliminated, I shall consider what I believe to be his principal contentions: (1) That my chart' shows the dividend-earnings ratios for 1936 and 1937 significantly below normal; that this implies the undistributed profits tax had the effect of causing retention of earnings; and that this extraordinary result casts doubt on the validity of the analysis. (2) That a better understanding of the effects of the tax is to be obtained from his time series chart of the Cowles Commission indexes of dividends and earnings; and that this chart shows the proportion of earnings distributed in 1936 and 1937 to have been exceptionally high. (3) That from the point of view of individual business men the effect of the tax upon aggregate earnings distribution isimmaterial; and that the law as drafted was inequitable, in that it placed the greatest burdens on small corporations without access to the capital market, and on corporations with large debts or deficits. First, the chart referred to (p. 344) was presented as a simple scatter diagram and not as a regression analysis precisely because I was unwilling to assign a normal percentage of earnings to be distributed at each level of business activity. While a general negative relationship between the dividend-earnings ratio and activity is familiar to all students of this problem, and is borne out by the scatter diagram, the myriad factors other than business activity which must influence corporate distribution decisions would make dependence on a mathematical equation relating these two factors alone foolhardy in the extreme. While I believe industrial activity to be the most important single variable influencing the dividend-earnings ratio, I recognize that only a third (31 per cent actually) of the variability in the latter may be accounted for by the former. If the years in which dividends exceeded earnings be omitted (a procedure Professor Guthmann advocates elsewhere in his note, but with which I cannot agree-cf. pp. 343, 345), the points for 1936 and 1937 lie even closer to the regression line; they are certainly neither significantly above nor below it.