ployment in the year. A decline in the volume of production for military purposes can be expected in this year. A substantial reduction in the production of equipment for land warfare is expected to be partially offset by an increase of material for naval and air warfare. It is also probable that there will be an increase in the production of civilian goods during this year, arising from reconstruction needs in Europe as well as from needs at home. No attempt was made to determine the general level of production. It may reasonably be expected to be lower than when production was required for two great theaters of war but higher than for any previous peacetime year.

Because of a lack of data from which the average length of haul of freight, the volume of railroad traffic, and the rate of productivity of railroad labor could be computed under the given conditions, the level of railroad employment was estimated at 1,550,000, a point approximately one-third of the way from the estimated level of peak wartime employment, 1,650,000, to the level expected in the first postwar year.

First postwar year

The first postwar year is likely to witness industrial activity at a level higher than any prewar year. A continuation of the high level trend of production (related to population) suggests an index of industrial production of approximately 144. This estimate is based on data for production in years of high industrial activity. (See table 28.) As the Federal Reserve Board index of industrial production, on which the above relationships were based, was not available for the period prior to 1919, a series was constructed for 1899–1919 in the same manner as by the Federal Reserve Board and spliced to its indexes. Since the ratios were obtained for years of high peacetime economic activity, and years reflecting average and depression conditions were omitted, they are not trend values as the term is commonly used but may be called a high-level trend. It is believed that such ratios express the favorable underlying conditions in the years covered.

TABLE 28.-Relationship of high-level industrial production and population under conditions of private enterprise economy, 1900-1950

1 Bureau of the Census for 1900-1940 figures; the 1950 figure is an estimate from the National Resources Planning Board, Estimates of Future Population of the United States, 1940-2000, p. 86, assuming low fertility, high mortality, and no net immigration, before adjustment for undercount of young children.

23-year centered averages for 1900 and 1910; 1919-20 average for 1920; 1928, 1929, 1930 average for 1930; 1940 average for 1940. The indexes for 1900 and 1910 are the National Bureau of Economic Research physical index of manufactures combined with production of minerals (as in the Federal Reserve Board index), spliced to the Federal Reserve Board index of industrial production.

* Estimated.

The indexes of production per 1,000 population indicate the growth of production due to factors other than that induced by the growth of population. The figures indicate that industrial production has grown substantially faster than population by an average of 28 percent of the 1900 volume of industrial production per 1,000 population per decade.

It was assumed that the nonpopulation factors would continue to operate at about the same rate as previously, so that the index of production per 1,000 popu lation would be 28 points higher in 1950 than in 1940. This increased index, in conjunction with the population estimate for 1950, suggests that a continuation of the high-level trend would mean that the index of industrial production would be 153 in 1950 and 144 in 1947. The latter was assumed to be the first postwar

year for the purpose of these calculations. Reconversion is expected to reduce the level of industrial activity from 144 to approximately 130. (See table 29.) Railroad freight traffic was estimated upon the basis of an observed relationship for the period 1932-40 between the volume of rail freight traffic for class I railroads, as measured in ton-miles, and the level of industrial production, as measured by the Federal Reserve Board index. Data for these years show that, in general, an increase of one point in industrial production resulted in an average of 2.22 billion ton-miles of rail traffic above 99.39 billion ton-miles. (See table 30.)

TABLE 29.-Estimated indexes of industrial production in the period of the adjustment to peace

1 Federal Reserve Board Index of Industrial Production, adjusted for seasonal variation. Annual indexes are averages for the 12 months indicated.

2 High-level trend value adjusted to allow for a 25-percent reduction in durable-goods production due to reconversion.

3 High-level industrial production trend value for 1947.

TABLE 30.-Regression equation of revenue ton-miles (class I railroads) on industrial production (FRB index), 1932-40

It was believed that ships would begin to resume their former importance in coastwise and intercoastal traffic soon after Japan is defeated and that a period of several months would elapse before their prewar position was resumed. An allowance was made from the railroads to haul approximately 40 billion ton-miles of freight traffic in addition to that indicated by the volume of production. This compares with an average volume of 200 billion ton-miles of such water-borne traffic for 1935-39. Since water routes are frequently more indirect than rail routes, the equivalent rail traffic load would be, roughly, two-thirds to threequarters of that by water. An additional allowance of 7 billion ton-miles was made for relief and rehabilitation traffic.

A factor tending to offset those mentioned above will be the large inventory of raw materials and semifinished products in the hands of fabricators. The value of inventories of raw materials and goods in process for war industries increased from 4 billion dollars in 1940 to 8.1 billion dollars in the third quarter of 1943, and for nonwar industries, from 3.4 billion dollars to 5.2 billion dollars. (Survey of Current Business, January 1944, p. 9). The increase in the physical volume of inventory is, roughly 50 percent. An allowance for the effect of large inventories was made on the assumption that manufacturers would have on hand approximately 1 month's extra supply at the 1940 level of raw materials and partially fabricated products. This would tend to reduce the volume of railroad traffic by approximately 12 billion ton-miles. The net allowance for freight traffic in addition to that indicated by the level of industrial production is thus 35 billion ton-miles. In applying the relationship observed between industrial production and rail freight traffic, and allowing for additional freight, an estimate of 423 billion ton-miles was obtained for the first postwar year. (See table 31.) This figure is 13 percent higher than that for 1940.

Railroad passenger traffic was estimated in a manner similar to that used for freight traffic. A relationship between population and passenger rail traffic was found to be fairly constant through 1920, but it did not hold for the last two decades. A rough relationship was found to exist, however, between industrial production and rail passenger traffic. Reduced to mathematical form, on the average there was found, for class I railroads, to be a change of 233 million railroad-passenger-miles, modified by a subtraction factor for each passing year, per unit change of the index of industrial production. (See table 32.)

TABLE 31.-Estimates of postwar freight traffic for class I railroads

1 Allowance for additional traffic resulting from delay in resumption of coast wise and intercoastal shipping and from relief and rehabilitation traffic, after adjustment for the effect on railroad traffic of large inventories of raw materials and semifinished products.

TABLE 32.-Estimates of postwar passenger traffic for class I railroads1

year-1935

+9,601 (0.6667) P-9,600. The exponent P to (0.6667) equals 5

For the purpose of estimating, the years are assumed to be 1947-51. The exponent P to (0.6667) equals year-1935.

5

3 Allowance for demobilization and high civilian passenger traffic due to large savings in the hands of consumers and remigration.

An adjustment to the estimated passenger traffic is necessary to allow for demobilization. The average rail travel required for each ex-serviceman demobilized from service abroad was assumed at 1,000 miles, and the average required from service in the United States was assumed at 250 miles. On this basis, 2.8 billion passenger-miles were allower for military needs. An additional allowance of 4.6 billion passenger-miles was made for civilian traffic arising from large savings in the hands of consumers, and remigration from war industry centers. The total passenger traffic for the first postwar year was estimated at nearly 32 billion passenger-miles, 34 percent higher than for 1940.

The level of railroad employment usually depends in part on the size of the traffic load and in part on the rate of productivity. Changing productivity in handling the traffic load was measured by computing the volume of traffic per man-hour. Traffic was measured by combining freight and passenger traffic into "traffic units." As computed in these estimates, the number of traffic units is equal to the number of passenger-miles multiplied by the ratio of the railroad expenses per passenger-mile to expenses per ton-mile plus the number of ton-miles. The total of traffic units for the first postwar year was estimated at nearly 575 billion. (See table 34.) This level is higher than that in any year in the period 1930-40, inclusive.

Table 33 presents, for class I railroads, the traffic unit performance, the number of man-hours paid for, and the number of traffic units per man-hour in past years. Since 1916, when satisfactory records were begun, the output of traffic units per man-hour has risen persistently. In only 6 years out of 27 did it fail to rise higher than any previously recorded figure. The average annual rate of increase during the period 1916-39 was 2.5 percent. The increase in productivity during 1940-43 was excluded from the computations because of the influence of the changed character of the traffic load and the improbability of the continuation of these rises in output per man-hour in the postwar period. Increasing the rate of productivity for 1939, 177 traffic units per man-hour, by 2.5 percent per year, to the first postwar year (assumed as 1947) provides an estimate of 216 traffic units per man-hour. Dividing the rail traffic for class I railroads by output per manhour yields the estimated number of man-hours required by class I railroads. Reducing the number of man-hours to average employment yields an estimate, for class I railroads, of approximately 1,107,000.

TABLE 33.-Traffic and productivity on class I railroads (excluding switching and terminal companies), 1916–43

? Computed by weighting the revenue passenger miles by the average ratio of expenses per passenger mile to expenses per ton mile for the period 1920-39 (4.74) and adding to the resulting figure the revenue ton-miles. All primary figures taken from ICC data.

Source: Interstate Commerce Commission. Years 1921-25 adjusted to allow for time paid for but not worked by freight train-and-engine service employees. Years 1916-20 adjusted to allow for time paid for but not worked by all employees.