that trend line. It fell with the depression of 1921; didn't quite recover to that trend line in the last part of the 1920's; fell sharply again in the early thirties; was barely restored to the trend line with the recovery programs of the 1930's. The recession of 1938 and 1939 lowered that ratio somewhat more and World War II brought farm income above the trend line. I want to point out, if the farm income estimate for this year materializes, the farm income ratio will then be below the trend line and we will, therefore, have the historical record from 1910 to 1950 revealing that in only two periods has farm income risen above the trend that might have been expected on the basis of the farmers' share in the national population, and it is for that reason that I say that, from here on, the farmers' prosperity will depend more and more on our ability to maintain full employment in the domestic markets.

TABLE 2.—Unemployment, the farmers' share in total relation of net income from farming to national income

1 Based on 1929-46 relation of Lebergott's percentages of time lost by unemployment to unemployment as percent of civilian labor force (Journal of American Statistical Association, March 1948).

The farmers' share of the national income derived from marketings is already lower than can be justified by the long-time decline in the farm share in the total population. For that and other reasons it is essential that industrial requirements for full employment are clearly visualized. It is for that reason I now wish to show you the current levels of steel production and capacity in relation to requirements over the next few years.

There has been and still is a great deal of disagreement regarding the adequacy of present steel capacity, the adequacy of plans for future expansion, and the relation of capacity production to profits, that I hope these simple illustrations will clarify.

Steel capacity is now at a record level. For the whole year of 1950, in view of current expansion programs, it will amount to about 100,000,000 ingot tons, exceeding the wartime peak of 96,000,000. But to say that steel capacity is now at an all-time high does not indicate its adequacy. The least that is required in judging an economic development of this sort is historical perspective. We need to examine that record capacity figure in relation to the long-time growth of the industry, to the growth of the national economy, and to current production requirements.

The capacity record as currently published by the American Iron and Steel Institute-our only source of information-is consistent back to the year 1925. The abrupt drop shown for that single year is merely a statistical drop marking a change in estimating. I did not include, Mr. Chairman, a copy of that record. It is abundantly available in many of the documents in these and other hearings, so I think perhaps the text will suffice or my comments will be sufficient without illustration.

It is necessary to make an allowance for that discrepancy to judge correctly the long-time growth of the industry. Making that adjustment, we obtain a record of steel capacity which arises from only 3.000.000 tons in the 1890's to 72,000,000 by 1929 and to 79,000,000 by 1932.

The effect of the deep depression of the 1930's and the abnormally low demand for steel was to check further expansion in total capacity until about 1938. By 1940 a moderate increase brought total capacity up to 81.6 million tons. But if the depressed demand of the 1930's had not interfered, we would have had about 92,000,000 tons in 1940 and therefore a larger base upon which to build wartime needs.

The wartime expansion induced and encouraged by the Federal Government was substantial, but it did not (at the peak in 1945) quite reach the level of at least 100,000,000 tons suggested by the longtime trend for that year. The resumption of expansion programs, after withdrawing obsolete capacity immediately after the war (in 1946), has now brought capacity to 100,000,000 tons and may carry the total a bit higher in the next few years. The main point to be observed, however, is that present capacity, had it followed the 40year trend from 1895 to 1935, would total at least 110,000,000 tons, or 10,000,000 tons greater than the present record figure. It is true that since 1940 steel capacity has kept pace with population growth, but that has not made up for the 10,000,000-ton shortage created by the depression of the 1930's. This is an important fact, generally overlooked, even by steel experts.

The record of steel production corroborates the conclusion that steel capacity of at least 100,000,000 tons would be about normal for 1950 and that an annual increase in that figure by perhaps 2,000,000 tons per year, as suggested by the long-time record, would be a normal trend of expansion for the 1950's.

The record of total demand for steel (domestic and for export) is a rising one. It tends to rise more rapidly than total population. Given full employment, it is likely to continue to expand. Forty years ago we were consuming steel at the rate of 600 pounds per person-20 years ago at the rate of about 900 pounds per person. By 1941, consumption had risen to over 1.200 pounds and could have gone higher had more capacity been available. Today, according to that experience, we would be consuming about 1,350 pounds, or about two-thirds of a ton per person. When I suggested 3 years ago that this figure would be appropriate for full employment conditions in 1950, it was generally considered as fantastic by those who failed to recognize the underlying character of the rising demand for steel under prosperity conditions, and by those who thought that the industry (in view of its sad experience in the 1930's) was overexpanded and should be cut back to a capacity of less than 80,000,000 tons.

96347-50-ser. 14, pt. 4a-51

The second chart is the one referred to by these figures. So, if you will be good enough, turn to the second chart entitled "Steel Capacity in the United States Since 1884," which is exhibit S-325 in Steel Exhibits, page 681.

The effect of the depression of the 1930's and the abnormally low demand for steel, and it might be helpful if at this point you will refer to the next chart which follows as exhibit S-326 in Steel Exhibits, page 682.

Spokesmen for the industry considered my projections so high that the American Iron and Steel Institute felt it necessary to publish a highly critical analysis of my statistical methods. That analysis presented two long-term trends which claimed to take into account "the basic curve characteristics yielded by mathematical computation, as well as the history of the steel industry and the fundamental nature of its products." At a time, namely, 1947-48, when steel consumption was averaging about 1,200 pounds per person and shortages and gray markets still prevailed, that analysis concluded thus:

The two curves

and you will find them indicated in the chart

mark, in this writer's judgment, the probable limits within which a long-term trend of the current period would lie, if determined 10 or 15 years from now by a good statistician familiar with the subject matter. For 1947 the calculations of annual steel consumption per capita indicated by curves A and B are approximately, and respectively, 830 and 985 pounds. For a population of 143,000,000, these amounts represent approximately 59,000,000 tons and 71,000,000 tons of demand. This compares with capacity of a little over 94,000,000 tons, and a current rate of production of approximately 85,000,000 tons.

The steel industry itself has already demonstrated that these conclusions erred on the low side. It is proceeding with an expansion program more nearly in line with my higher estimates of capacity and production requirements. When demand for steel fell off in 1949, production declined to an annual rate of 78,000,000 tons (substantially above the range of 59 to 71 million tons in the quotation above). The industry is now operating its 100,000,000-ton capacity at about 100 percent, supplying about 1,280 pounds per person. Demand in the second quarter of 1950, judging by trade reports, is in excess of present capacity.

This condition of demand, it is important to observe, exists in spite of the fact that present employment conditions are about 2,000,000 short of practical full employment. If we had actual full employment, we could today be consuming about 1,350 pounds of steel per person. This would be even more closely in line with what my projections of 3 years ago indicated.

The lower estimates shown in the chart, Three Estimates of Steel Production per Capita Required for Full Employment, United States, were presented by Mr. Wilfred Sykes, president of the Inland Steel Co., June 19, 1947, before the Steel Subcommittee of the Special Committee to Study Problems of American Small Business, and by Mr. Bradford B. Smith, economist with the United States Steel Corp. in an undated pamphlet entitled "America's Steel Capacity, What It Is, What It Does," published by the American Iron and Steel Institute, I believe early in 1948.

If now we can more readily accept a per capita consumption of 1,350 pounds (or two-thirds of a ton) as about normal for full em

ployment, it is clear that for the present population of 152.000.000 we need a normal full-employment production of about 100,000,000 tons. That, however, would be pressing present capacity to the limit Capacity normally should exceed production required for conditions of full employment. This would seem to be in harmony with the following view introduced by the United States Steel Corp. into the record of the Temporary National Economic Committee (exibit 149. p. 149):

Ingot capacity, an accepted basis for determining rates of operations, reflects roughly operations of finishing capacities. Even in periods of peak demand, orders are not distributed among products in such a way as to make possible filmtilization of all finishing facilities. In practice, therefore, operations probaby would never be maintained at 100 percent of either inget or finishing capacity because of lack of coordination between demand and capacity for various products. Production might, therefore, be expected to run below capacity even at the peak of the cycle.

In times of emergency, or under the pressure of extraordinary demands in the industry, it might occasionally be possible to attain an operating rate slightly in excess of 100 percent of rated capacity for short periods by bringing into operation obsolete facilities, lengthening the workweek, eliminating holidays, or by other meals.

These considerations, as well as the desirability of avoiding marketprice advances either black or gray, during periods of full employment, suggest that for a Lormal production of 100,000,000 tons, normal capacity could be 110.000.000 tons or more.

Estimates of capacity and production requirements for full employment over the next 10 years will, of course, differ according to the estimates of population growth in national production of goods and services requiring steel, the supply of ore and other raw materials, and many other considerations, technological, international, and strategic. It is probably sufficient for the present to suggest that if by 1960 per capita consumption should continue to expand to, say, three-fourths of a ton per person, and if total population should increase at the rate of only 1 percent per year, say, to 165,000,000, we would need a total production of around 125,000,000 tons per year and a capacity to provide the necessary "elbow room" of at least 135.000.000 tons.

One might go one step further in industrial arithmetic and note what expansion possibilities of this sort might mean for existing firms or newcomers to the steel industry. Assuming no new firms entering the field, the possible expansion over the next 10 years could mean an increase in total volume for existing firms of about a third. In the past the largest corporation, it would seem, has followed a policy of expansion closely approximating the growth of population. Expansion in excess of population growth has been supplied by companies other than United States Steel subsidiaries. If this tendency were to prevail over the next decade, it would mean that the expansion opportunities for the other two-thirds of the industry (in terms of capacity) would be proportionately greater by 1960 than the 35 percent increase indicated for the industry as a whole. In such an expansion there would seem to be ample room for new steel companies. These speculations suggest two other observations, namely, that such an expansion would of course be an advantage to the industry in terms of profits for a rising volume and mounting assets, and of

considerable help in the long-run need for industrial opportunities for underemployed persons now engaged in agriculture.

For the steel industry there are profits in sight as long as demand exceeds 40 percent of capacity. If we consider the 1932 depression as nontypical, steel demand over the years is quite likely to be well above 40 percent.

The following table 3 gives the recent profit experience for the industry as a whole. It shows that when production is 80 to 90 percent of capacity, the industry (as represented by 11 firms) earns 8 to 11 percent on investment. At 40 percent of capacity, profits are on the average reduced to zero.

TABLE 3.-Production as percent of capacity and percent earned on investment

1 American Iron and Steel Institute, Total Industry.

2 Federal Trade Commission, Eleven Principal Steel Companies.

NOTE. See chart United States Production of Steel, Percent of Capacity, and Percent Earned on Investment by Eleven Principal Companies which appears as exhibit S-327 in Steel Exhibits, p. 682.

Using a consistent method of computing profits (no allowance in 1948 and 1949 for accelerated depreciation), the Federal Trade Commission finds that profits averaged about 22 to 3 percentage points in excess of the earnings for comparable rates of operation during the period 1932 to 1941. This suggests that for 1948 and 1949, at least, the break-even point (about 40 percent during the 1930's) has not risen as is generally assumed. It has been argued that the break-even point for the steel industry is far above 40 percent of capacity, but the 1948-49 experience does not seem to support that view. In considering the effect of further expansion on profits, it is important to note that the relation of volume to profits remained practically unchanged, in spite of the marked changes in production costs and sales prices that took place between 1932 and 1949.

In the past 30 years covered by the Federal Trade Commission profit estimates, and for a much longer period covered by the industry's records, there have been only 5 years of production at 40 percent of capacity or less, and profits at zero or less; namely, 1931-34 and 1938. This generalization holds also for the United States Steel Corp., data for which appear in table 4 and the accompanying chart (United States Steel Corp. Production as Percent of Capacity and Percent Earned on Investment). Profits in 1948 and 1949 were in line with the past relation of production to profits. In only 6 years of its entire history of nearly 50 years has the corporation operated at 40 percent. of capacity or less (1931-35 and 1938) and with profits at zero or less.

1 This chart appears as exhibit S-328 in Steel Exhibits, p. 683.