productive capacity, industrial production, and steel requirements
TRANSCRIPT
STEEL REQUIREMENTS
trial production, PT, steel requirements will tend to be lowerthe higher the bench mark value is, and they will, in general,also tend to be lower the lower PT is relative to the bench mark
The derivation of Formula G holds, not only for the casen = 2, but, in general, for any number of industries.
In addition to the "product-mix" effect in the case where cy-clical sensitivity is correlated with unit requirements of steel,another factor contributes to the cyclical sensitivity of steel con-sumption: working-inventory requirements. Working inventoriesare built up or depleted according to whether a given level ofproduction is reached on the upswing or on the downturn. Thisis true for every component industry and ipso facto for totalindustrial production itself. When, in the gradual long-run ex-pansion of production, a level of 90 on the Federal Reserve Indexwas reached from below—as in 1925—working-inventory require-ments expanded, as reflected by a rate of 564,000 tons of steelper point on the Index. But when total industrial production,after having reached a high of 110, fell off to a level of 91 onthe Index—as in 1930—inventories, previously geared to higherproduction levels became excessive as working inventories andthus depressed the demand for new steel below the level indi-cated by the long-run expansion of production. Steel require-ments in 1930 amounted to 501,000 tons per point on the Index.
Thus the working-inventory factor alone would be able tocause both a moderate rate of increase in steel consumption ina gradual long-run expansion of industrial activity and a muchsharper backsliding of steel demand during a recession of pro-duction. This is shown by the pattern of successive cycle swingsin Chart 1. Since two factors are operative, one might ask whethertheir effects can be separated. The results of a test introducingthe annual rate of change in production as an additional factorin the analysis are reported in Section F-S.
First, however, the general hypothesis embodied in Formula Cwill be tested.
C. THE ASSOCIATED SECULAR FUNCTIONOF INDUSTRIAL PRODUCTION
In order to test Formula C, it is necessary to develop a suitablemeasure for the "associated secular function" of the index ofindustrial production corresponding to the variable of this
240
S
STEEL REQUIREMENTS
formula. This series must, in some sense, represent a measure of"normal" production, so that the relation of actual industrialproduction to it will provide a reasonable measure of cyclicalslack or strain.
Such a series could be derived in various ways. One could,for instance, apply directly the specific technique used by Modi-gliani and define the associated series in the year t, (t), as thehighest level of total industrial production preceding the year t.Another simple method has been used by A. C. Hart.8 Accordingto this method, the associated function would consist of a timeseries of annual points including the successive peaks of industrialproduction and points interpolated annually between them lead-ing in geometric progression from one peak to the next.
While any of these methods would probably yield a reasonableapproximation, a somewhat different approach will be used here,which, at least conceptually, appears more suitable for the pur-pose on hand. The associated function in any given year will beexpressed as an estimate of productive capacity for that year,measured—like production itself—in points of the Index; in otherwords, as a series measuring the highest level that total industrialproduction could reach on the basis of productive capacity inexistence in that year.
The relation of production to productive capacity is a reason-able over-all measure of cyclical position. Indeed, the differencebetween actual production and productive capacity seems to bea very good operational measure of the notion of cyclical slackor strain. In addition, such a relation may be expected to indi-cate the effect of cyclical strain on the activity of producers'equipment industries. These industries are heavy users of steel,the main raw material for producers' equipment. Such a rela-tionship for manufacturing and mining equipment is illustratedin Chart A-i and Table B-2 (see Appendix B).
The task of developing a measure for over-all productive ca-pacity obviously presents some serious problems, both concep-tually and statistically. Some of these problems are mentionedin Appendix A, in which an estimate of productive capacityfor every year from 1919 to date is developed. This estimate wasconstructed on the basis of yearly data on the flow of equipmentinstalled and on yearly estimates of retirements of existing ca-
S A. G. Hart, Money, Debt, and Economic Activity (Prentice-Hall, 1948),pp. 260ff.
241
STEEL REQUIREMENTS
pacity.° The estimate represents no more than a rather crude,though useful, over-all approximation of the theoretically rele-vant concepts. There has been considerable interest in the gen-eral problem of developing, both conceptually and statistically,suitable measures of productive capacity. Such measures have anumber of important potential applications, of which the presentis but one. It is to be hoped that the work now in progress willyield more refined methods of estimation and improved estimatesfor the past.
D. STATISTICAL TESTS OFTHE HYPOTHESIS
In order to test the hypothesis and to estimate the coefficients ofthe general formula G, total steel production ST is correlatedwith total industrial production PT (Federal Reserve Index) andwith the measure of productive capacity PCAP. The conditions ofthe period 1919-40 are then summarized as follows:Formula II
ST= k + pPT — q•PcAp= 15.40 + 0.857PT — 0.4OSPcAP
[5.48] [0.044] [0.052] R = 0.969
The resulting parameters are obviously quite favorable to the hy-pothesis. The coefficient of PCAP is negative, as expected, and ishighly significant by standard statistical tests, being 7.75 times aslarge as its standard error. The multiple correlation coefficient risesto 0.969, as against a simple correlation of 0.897 for Formula I.
In order to see fully the implications of Formula II, it will beuseful to perform certain algebraic transformations. In the firstplace, it must be remembered that the variable PCAP denotes themaximum level consistent with existing capacity. In general,however, a level of production equal to capacity, that is, a 100percent rate of use, can hardly be considered normal. Ideally
It is likely that installation of equipment is highly sensitive cyclically.Thus, when the surviving annual equipment installations of preceding years(or their capacity equivalents) are cumulated, it must be expected that theseries of productive capacity will show an echo effect, that is, ripples of thedamped and lagged effect of the business cycle. As far as these ripples arestill present, productive capacity is only an approximation of the "associatedsecular function." The series on productive capacity was originally devel-oped for the Econometric Institute, Inc. It is described in Appendix A. Itforms part of the copyrighted service of the Institute, and is presented hereby permission of its president, C. F. Roos.
242
STEEL REQUIREMENTS
the "normal" rate of use might be defined as that rate at whichfirms, in the aggregate, do not wish either to expand or to con-tract their capacity. In general, such a rate of use will be sig-nificantly less than 100 percent of POAP, especially since the meas-ure of capacity is a ceiling for the production during any monthof the year and enough slack capacity must be allowed to takecare of seasonal fluctuations in production. We assume that 85percent of capacity represents a normal rate of use. In fact, thisassumption is implicit in the method of estimating the capacityseries.
On the basis of this assumption, the "normal level of produc-tion" for the year t, or PN ( t) would not equal 100 percent ofPCAp(t), but only 0.85 PcAp(t) or
PN=PN(t) =0.85PCAP(t) (10)In the following reformulation of Formula II, PN, the "normal"
level, will be used as the "associated secular function." Thequalffications which apply to PCAP (see footnote 9) apply alsoto PN. Production, PT, will not only fall short of capacity, PCAP,but more often than not, of the normal rate of production,PN — 0.85 PCAP. This was the case in 15 out of 22 prewar ob-servations. Only in 5 cases did PT exceed PN. Thus it appearedadvisable to introduce the cyclical position of production not asa cyclical PN, but as a cyclical slack, PN — PT. Itshould be noted that the substitution of PN for PCAP, while log-ically desirable, does not affect the results substantively.
Substituting (10) into the general formula C and into theformula with numerical parameters (II) leads to
ST=k +pPr — (q/0.85). (0.85PcAp)= k + p PT — q' PN (C-a)
15.40 + 0.857PT — 0.474 PN (Il-a)These formulae may be rewritten in the following forms, which
will be used in the following discussion:ST= k +(p—q')PT— q'(PN—PT) (C-b)
= 15.40+ 0.S8SPT—0.474(PN—PT) (Il-b)or also:
ST= k — p(PN—PT)+(p—q').PN (C-c)=l5.4O—O.857(PN—PT) +0.383 •PN (Il-c)
• From the expression on the right-hand side of Formula Il-a,243
STEEL REQUIREMENTS
it is apparent that as long as PN remains constant, ST will tendto change by 857,000 tons per point of the Federal Reserve Index.In the short run, PN will be relatively stable compared withIndeed, especially if fluctuates below PN (this has been calleda "cyclical fluctuation" in production), PN is likely to changevery little, if at all. Hence, given a stable value the relationbetween steel output and industrial production during any givencycle may be approximately described by
ST= (k— q' PN) + PPT (G-d)= (15.40 — O.474PN) + 0.857 PT (II-d)
This is a line with considerably steeper slope than the slopeof 0.697 million tons per point on the Federal Reserve Indexindicated by the "hybrid" Formula I. It is indicated by thedashed-dotted line on Chart 1. As PN changes_that is, increasesfrom one cycle to the next—this line will shift to the right asindicated by the configurations on Chart 1 for the years 1919-22,1929-36, and 1937-39.
Alternative statements of Formula G—(G-b) and (G-c)---aredesigned to bring out the long-run aspects of the demand forsteel more clearly.
A long-run change in the demand for steel may now be opera-tionally defined as a change in demand' S, associated with achange in the over-all level of production P, when enough timehas been allowed for the level of capacity to adjust itself to thenew level of production. Clearly this definition of "long run" isbasically consistent with the traditional Marshallian notion. Bymeans of Formula Il-b, the long-run change in the demand forsteel which is associated with a change in production can becomputed. Suppose the starting point is the year 0 with a levelof production PT (0). Suppose now the level of capacity is fullyadjusted; then
PT(O)—PN(O)=Oand therefore
ST(0) = 15.40 + 0.383 PT(0)Suppose that production rises to a new level
PT(1) = PT(0) +If, as has been assumed, capacity is fully adjusted to this newlevel, or
PT(l) —PN(1) =0244
STEEL REQUIREMENTS
and henceST(1) = 15.40 + 0.383PT(1)
then the change in demand will be given by
Thus Formula II implies that, in the long run, the demand forsteel would tend to change only 0.383 million tons per point ofthe production index, as contrasted with a rate of 0.857 milliontons per point of the production index in the short run and a rateof 0.697 million tons on the production index in the hybrid For-mula I. The equation obtained for the vanishing cyclical slack(PN — PT) = 0 is
ST=k+(p—q')PT=k+ (p—q')PN=k+nPN= 15.40 + 0.383 PN (G-e)
This equation may thus be considered an estimate of the long-run relation between the demand for steel and industrial produc-tion. The coefficient n = p — q' is an estimate of the long-runmarginal steel requirements per unit of change in the produc-tion index. The long-run relation is indicated by the slope of thedashed line in Chart 1 and by column 1 of Table B-4 (see Ap-pendix B).
E. EXTRAPOLATION TESTSThe introduction of the variable PN in addition to PT resulted ina substantial reduction in the variance which was not explainedby Formula I. A more significant test of Formulae I and II, be-cause the parameters are based on prewar observations only, isto use these equations for an estimate of the demand for steelfor the postwar period, say, around 1950, when production wasat much higher levels. Since the change in production between1940 and 1950 can be regarded largely as a long-run change,Formula II should be expected to yield somewhat more accurateresults than Formula I. At the same time a margin of error con-siderably larger than in the period of observation must be ex-pected, since the stretch of years between 1940 and 1950 issubstantial, and some changes in the basic relation between pro-duction and steel requirements might have occurred.
If Formula Ills used to estimate the demand for steel at the1950 level of production (200), it is also necessary to specifythe assumption about the associated secular function, or its ap-
245
STEEL REQUIREMENTS
proximation PN. An assumption of no cyclical slack, that is, com-plete adjustment of capacity to this production level, yields theestimate ST ( 1950) = 15.40 + 0.383 ( 200) = 92 million net tons ofsteel. The actual level of production in 1950 was 96.8 million nettons, so that this long-run projection represents an underestimateof about 5 percent. For the year 1951, the underestimate is ofthe same order of magnitude, with an estimated level of 99.6million tons and an actual level of steel output of 105.1 milliontons. This represents a striking improvement over Formula I, theuse of which leads to overestimates in the order of 25 and 30percent, respectively, for these two years. (See Table B-3, col-umn 2, in Appendix B.) With a value for PT(1948) of 192, thelong-run projection of steel demand is 88.9 million net tons;with a value for PT (1949) of 176, it turns out to be 82.8 millionnet tons. These estimates compare with actual levels of steeloutput of ST( 1948) = 88.6 million net tons and 1949) = 78million net tons. In every case the error is in the order of 5 per-cent or less. These results can also be obtained from Chart 1, byreading the projected long-run steel demand at each level oftotal industrial production from the dashed line showing thislong-term projection. The vertical excess of actual steel outputfrom each of these levels gives the extent of the deviation inmillions of net tons.
Actually it is possible to develop some supplementary, thoughtentative, information on PN for the postwar period. It is shownin Table B-2, column 4. Substituting these values in Formula IIleads to a more accurate extrapolation, shown in column S ofTable B-S and in column 1 of Table B-4. For the five-yearperiod 1947-51, actual steel production averaged 90.7 million nettons, estimated steel production derived by Formula II amountedto 92.5 million net tons. The absolute average error is 7.4 millionnet tons.
At this point it may suffice to point out that Formula II per-forms far better than Formula I and that its use for long-runprojections is reasonably satisfactory. Before further extrapola-tion tests are undertaken, some refinements of the hypothesismust be introduced.
246
STEEL REQUIREMENTS
F. SOME REFINEMENTS OFTHE HYPOTHESIS
1. Changes in rates of consumption of steel withinthe period of observation, 1919-40
The period 19 19-40 is a long one, in the course of which somechanges in the underlying relation between steel output andtotal industrial production might have occurred. To test thishypothesis, two sets of coefficients, p and n, one for the period1919-29, the other for the period 1930-40, were simultaneouslyreestimated. We call the analysis of total production of steel foringots and castings with different levels of p and n Formula III.To distinguish the reestimated coefficients and the constant fromthose derived for the period as a whole in Formulae G-a—e, thenumber of the formula is introduced as a subscript of the param-eter, e.g., pu, viiI, etc. refer to the coefficients of total industrialproduction in the general formula C-a or to that of the cyclicalslack (PN — PT) in Formula G-c, while n11, nm refer to the co-efficient of PT in (C-b). Finally k11, k111 refer to the constantterm, which covers the entire period 1919-40 in each of theseformulae.
A comparison of the coefficient for the two separate periods(see Table 1) reveals a substantial decline from the twenties tothe thirties in the short-term rate of steel consumption and a lesspronounced decline in the long-term rate n111. There are goodreasons for these declines in steel requirements per unit of pro-duction. It is probable, in the short term, that improvements inthe control of flow and the stocking up of raw materials for oper-ating inventories took place. The cooperation of some steel pro-ducers with automobile manufacturers in an effort to tighten upsteel delivery schedules is one of several examples. In the longrun, use of better-grade steel and more elaborate end-use speci-fications reduced the requirement per unit of output. How closelyFormula III fits the data can be seen from Chart 2. In this chart,actual steel production is shown as a heavy solid line, while steelproduction as estimated from Formula III is represented by athin line. (For all other lines see Section H.)
2. Demand for steel for the domestic marketNot all finished steel produced from the output of steel ingots
247
STEEL REQUIREMENTS
TABLE 1COMPARISON OF THREE ALTERNATIVE ESTIMATES OF COEFFICIENTSFOR THE GENERAL FORMULA S = k + n — p. (PN — PT)
Period
Parameter 1919-40 1919-28 1929 1930 1931-40
Constantk11k111
15.40(5.48) — — —5.575(10.7669) — —0.744(10.4623) — — —
—
nu
niv
Coefficient for the normal rate of production0.383(0.076) — — —
— 0.5073(0.2372) 0.4922 0.4719— 0.5225(0.2216) 0.4986 0.4668
—0.4568(0.1227)0.4429(0.1191)
pu
puipuv
Coefficient for the c!,clical strain, PN —0.857(0.044) — — —
— 0.9916(0.1491) 0.9323 0.8532— 1.0705(0.1448) 0.9413 0.7690
0.7939(0.0486)0.6398(0.0471)
Note: Standard errors of regression coeflicients are given in parentheses. The transition be-tween the twenties and the thirties has been derived heuristically. Various types of splicingwere tested and the one with the least error of estimate selected. This particular form ofing does not jump immediately from higher level of the parameters in the twenties to thelower level of the parameters for the thirties, but provides for two intermediate stages in thetransition: the year 1928 is the last year of the higher level, the year 1931 the first year of thelower level for the parameters. In descending from the higher level, the year 1929 is placed 0.3of the total jump below the higher level; 1930 is placed 0.3 of the total jump above the lowerlevel for the parameters.
and castings is consumed in the domestic market.10 While UnitedStates business cycles may affect the rest of the world, factorsother than industrial production in the United States also havebeen determinants of foreign demand for American steel. Forinstance, the shortage of steel in other countries after each ofthe two world wars caused exports of steel to rise more sharplythan domestic consumption. The proportion of United States do-mestic raw steel production ultimately shipped abroad as finishedsteel has varied from 3 percent to 18 percent during the last thirtyyears. For this reason the coefficients of Formula III were re-estimated and only the domestic suppiy of steel was used as adependent variable.
10 In order to estimate this portion, it was necessary to translate theexports of rolled steel (Metal Statistics, 1951 [American Metal Market,1951], p. 260) into the equivalent steel for ingots and castings. The con-version factors were derived by comparing total rolled steel for sale . iithtotal ingot production, assuming a one-month lag of rolled steel shipmentsafter the production of ingots.
248
STEEL REQUIREMENTS
CHART 2
Steel Output and Capacity and Estimated Demandfor Steel at Various Levels of Industrial Production
249
Millions of net tons
STEEL REQUIREMENTS
The estimates thus obtained are denoted, niv, and pzv.They are compared in Table 1 with the various estimates dis-cussed previously.
An interesting result of the successive refinements of the anal-ysis is the progressive fall in the constant term from 15.40 millionnet tons for Formula II to less than 1 million 'net tons forFormula IV.
It should be noted that the original hypothesis, as expressedin Formula C, lacked a constant term. There is no reason to ex-pect that this condition would be exactly satisfied by a regressionequation fitted to the data. The measure of total production isnot only directly representative of manufacturing and miningindustries consuming steel, but it acts also as an indirect measureof the variations in the demand of nonmanufacturing steel con-sumers, such as railroads or public utilities. Nonetheless the con-stant term of 15.40 million net tons was uncomfortably large. Ithad to absorb the effect of averaging over the entire period 1919-40 in the determination of the value of nn. Inspection of nutindicates that while the rate for the thirties was 10 percent be-low that of the twenties, it was still 16 percent above the "av-erage" rate of n11 for the entire period. The latter is, therefore,not a pure average of the two rates 0.5073 and 0.4568, whichought to be around 0.48, but contains also the slurring effect ofmoving from the higher long-run relation to the lower long-runrelation for the later years. The difference in the amount ofvariation, explained by an average of the long-run parametersof Formula III and by that of Formula H, appears as a differenceof about 10 million net tons between the constants of the twoformulae.
Formula IV not only confirms the decline in the rate of steelrequirements, but suggests that the difference in this respectbetween the thirties and the twenties is even greater than thatindicated by Formula III: for p, the drop amounts to 20 percentin the case of Formula III and 40 percent in the case of FormulaIV; for n to 10 percent and 15 percent. An analysis of these shiftsfrom the viewpoint of technology would be desirable. TheFormula III estimates of 95.9 million net tons for 1950 and 104.6million net tons for 1951 come closer to the actual performanceof 96.8 and 105.1 million net tons than the Formula IV estimatesof 92.8 and 102.4 million net tons.
The mean square deviation for the period 1947-51 amounts to
250
- w
STEEL REQUIREMENTS
7.6 million net tons in the case of Formula IV, 8.3 million nettons in the case of Formula III, 8.5 million net tons in the caseof Formula II, and to 27.8 million net tons in the case of FormulaI.
As stated earlier, it should be noted that there are other seg-ments of steel consumption in addition to exports which are notmeasured by the Federal Reserve Index of Industrial Production.A certain portion of steel is shipped directly to railroads, utilities,waterworks, farmers, government, etc., and thus does not passthrough any manufacturing stage outside the steel industry. Tothis extent the parameter of Formula IV contains an indirectweight for such steel consumers.
3. The rate of change of industrial production as a factorin the demand for steel
In the theoretical discussion of the basic approach to the generalFormula G, the product-mix change was given as one reason whythe cyclical position of industrial production ought to be con-sidered as well as the rate of production. The working-inventoryrequirements for steel were given as another factor. Working-inventory changes resulting from changes in level of produc-tion are not necessarily cyclical in nature. It may, therefore,be asked whether we can substitute the change in the level ofproduction for the cyclical slack of production and reach anequally satisfactory explanation of steel requirements. Will theintroduction of the change in the level of production as a thirdsupplementary variable contribute to a sharpening of the analysisand to a separation of the "product-mix" from the "working-in-ventory" effect?
In order to obviate, at least partially, the inadequacy of annualdata on production changes as indicators of inventory move-ments, two different forms of the annual change in productionhave been tested, both by use of Formula IL
a. The change from the last year—which amounts roughly to thechange during the preceding fiscal year—has shown a positiveeffect. At maximum it amounted in 1938 to 4.8 million tons, ifused as a substitute for the cyclical slack in production, and to1.8 million tons, if used as a supplementary third variable. Ingeneral, a change of one point on the production index is ac-companied by a change in steel output of nearly 200,000 tonsand 75,000 tons, respectively. The reduction in unexplained varia-
251
STEEL REQUIREMENTS
tion resulting from the introduction of the change during thefiscal year is, however, insignificant.
b. The change during the calendar year—which points furtherinto the future—has a negative effect of 46,000 tons per point onthe production index, if used as a substitute, and of 39,000 tonsif used as a supplementary variable. Its introduction scarcelyreduces the unexplained variation.
If an analysis of inventory changes based only on annual datacan be considered fair evidence, the effect of these annualchanges is minor, or insignificant, and an ex post adjustment,rather than an immediate adjustment or anticipation.
G. EXTRAPOLATIVE VALUE
1. Wartime conditionsThe formulae cannot be applied to steel requirements under war-time conditions: POAP is a measure of peacetime civilian Ca-pacity. Total military and civilian production exceeded peace-time capacity in the two main war years 1943 and 1944 by about80 points. The measure of production expanded by 137 pointsover 1935-39, while peacetime civilian capacity, which excludesthe capacity of ordnance and similar plants, advanced only 28points. Peacetime capacity also does not allow for multiple-shiftuse of a plant if normally it is used only in single-shift operations.It also excludes the obsolete or obsolescent stand-by equipmentreturned temporarily to operation under emergency conditions."The development of a properly calibrated measure of wartimeadditions to capacity, which would take all the distortions intoaccount and allow for the shifts in the composition of produc-tion, has not been undertaken. A proper test of the formulae byextrapolation is, therefore, confined to the postwar period.
"The war period was characterized by a rapid expansion of the ma-chinery and transportation equipment industries. This can be explained, inpart, by the greater amount of fabrication and inspection actually requiredfor the output of combat material, and, in part, statistically, by the use oflabor input as a measure of output. During the prewar period, 1919-40, themetal-fabricating segment of the production index amounted to 1.1 timesits basic metal segment plus 1 point on the index. The equivalent for thepostwar period was about 11 points higher than the prewar relation. Duringthe height of the war, 1943 and 1944, the metal-fabricating segment actuallyrose 90 points above the prewar relation. In 1951, it rose 4 points above thepostwar relation.
252
STEEL REQUIREMENTS
2. The postwar periodFormulae II, III, and IV overestimate steel requirements in thefirst few years of the postwar period. This result emphasizes thetemporary effect of wartime demand and suggests that steel re-quirements per unit of production may have declined. On theother hand, these overestimates decline from year to year untilthey change into underestimates either in 1949 or in 1950. Sucha pattern would point to the existence of real steel shortages,especially in the strike years 1946 and 1949. On the whole, itseems not unlikely that the truth is somewhere between the twoextremes: part of the overestimate from 1946 to 1949 may reflecta true shortage of steel, while part of it may reflect a genuineoverestimate. Formula IV would result in a gross overestimateof about 40.7 million net tons for the three years 1946 to 1948.If—by way of an illustration—this gross overestimate is formallysegregated into a "genuine" steel shortage of 25 million net tonsand a "genuine" overestimate of about 15 million net tons, theeffect on the parameters of Formula III or IV can be expressedas a drop of 8 percent in the short-term parameter, as comparedwith a drop of 40 percent in the same parameter between thetwenties and the thirties; and a drop of 8 percent in the long-run parameter niv as compared with a drop of 15 percent be-tween the twenties and the thirties. Thus the formal assumptionof a steel shortage of about 25 million net tons during 1946-48would suggest that a further drop of unit requirements for steelhas occurred, amounting to about one-fifth of that which char-acterized the transition from the twenties to the thirties. Furtherevidence will be needed before a definite statement can be made.
H. STEEL CAPACITY AND STEEL REQUIREMENTSFOR FULL-CAPACITY PRODUCTION
One interesting application of Formula G is its use in estimating"steel requirements at capacity operations," that is, steel require-ments for a level of industrial production equal to capacity. Thisrepresents an estimate of the maximum steel requirements in anygiven year.
The original formula
S(PT,PCAP)=JC+P•PT—q.PCAP258
STEEL REQUIREMENTS
is a convenient startftig point for estimating "steel requirementsat capacity operatiàns." Substitution of PT = PCAP leads to
S (PCAP) = S (PCAP, POAP) = k + (p — q ) POAP
(n+q'—0.85q')PCAP=k+[n+0.15(p—n)]PCAP (11)
Using the values n1ij = 0.5073 to 0.4568 and = 0.9916 to0.7989, the coefficients of PCAP assume the following values: for1919-28, 0.580; for 1929, 0.558; for 1930, 0,529; and for 1931-40,0.507.
These estimates of "steel requirements at capacity operations"are shown as a time series in Chart 2 and compared with steelcapacity, actual and estimated steel production, and steel re-quirements at the "normal" level of production (that is, 85 per-cent of capacity). "Steel requirements" according to Formulae II,III, and IV are shown for the "normal" level of production incolumns 1, 2, and 3 and for capacity industrial operations incolumns 4 and 5 of Table B-4 (see Appendix B).
An inspection of Chart 2 reveals the relative closeness of actualsteel capacity and "steel requirements at capacity operations" asestimated from Formula III. During the thirties—as a result ofthe decline and the slow recovery of industrial capacity—"steelrequirements at capacity operations" fell below actual steel Ca-pacity, which remained relatively constant. If the rate of short-term, or output-determined, and the rate of long-term, or ca-pacity-determined, steel requirements are applied to the postwarperiod without any further reduction, it appears that industrialcapacity has expanded much faster than steel capacity. During1946, when existing steel capacity dropped nearly 5 percent asa result of writing off obsolete facilities, "steel requirements atcapacity operations" started to exceed actual steel capacity. Infact, steel capacity fell so far behind that it was only 2 milliontons above the steel requirements for the normal level of activity(85 percent of capacity) for the year 1950, thus leaving littleleeway for the expansion of activity beyond this level. Since steelcapacity was so much lower in 1950 than the steel requirementsindicated by estimates of peacetime production capacity, onemight expect it to have had a limiting effect on the expansion ofindustrial production in recent years.
254
STEEL REQUIREMENTS
I. STEEL CAPACITY AS A LIMIT TO THETOTAL LEVEL OF INDUSTRIAL PRODUCTION
An estimate of the ceiling imposed on the possible level of totalindustrial production by the available capacity for producingsteel can be derived by inverting Formula G and solving it fortotal industrial production From
ST=k+P•PT—q•PCAP (12)
it follows thatPT = (q/p)PCAP — (l/p) (S — k) (13)
Substituting first q = 0.85. q' and then q' = p — n into theformula leads toFormula V
PT = 0.85(1 — n/p)PCAP + (1/p)(S — k)
Formula V is to be regarded as measuring the amount ofindustrial production that can be supported by a given supplyof steel, S. Introduction of estimates of the parameters n, p, andk and of the level of steel capacity SCAP in Formula V yields anestimate of the maximum level of industrial production con-sistent with the available capacity to produce steel. If the valuesassumed by the coefficients for the period after 1930 (niir =0.4568, pin = 0.7939, and k111 = 5.575) are used, and if the val-ues prevailing in 1952 are approximated by POAP (1952) = 286and SCAP (1952) = 111.5 million net tons, the production po-tential can be estimated at
0.4246 .243 + 1.2596(111.5 — 0.0 — 5.6) = 237 points (14)The ceiling for total industrial production is determined by thesteel supply for the domestic market, or 111.5 — 3.8 million nettons, rather than the total steel supply. Using the comparableparameters of Formula IV for the period after 1930 results in
0.3078•243+ 1.5630(111.5—3.8—0.7) =239 (15)Introducing the tentative reduction of steel requirements de-veloped in Section G-2 would increase the industrial productionpotential to about 245 points, or 6 points above the results basedon 1930-40 parameters.
J. SUMMARYEarlier investigations have demonstrated that price alone is nota major determinant of changes in steel demand. More and more
255
STEEL REQUIREMENTS
emphasis has been placed upon the shift of the demand functionitself. Steel constitutes a unique raw material for many durablegoods. The area where substitutions are technologically or eco-nomically feasible comprises only a minor fraction of aggregatesteel demand. This being the case, how will the steel consumerwho wants to stay in business react to an increase in steel prices?He will ask whether it is possible to shift the whole cost increaseto his customers without impairing the size of the market for hisproduct and he will consider whether his profits permit him toabsorb the price increase. Only in a few cases will he find itpossible to use a substitute or to reduce his unit demand for steelon short notice. In the long run, one must expect steel savingsto result more from the continual striving for efficiency and gen-eral cost reduction rather than in response to isolated shocks ofincreases in the price of steel. Thus, using the language of theclassical demand and supply analysis, it is necessary to explainthe shifts in the demand function itself. The extent of the shiftsof the demand function will make shifts along the demand curvedue to price changes appear unimportant.
An attempt at explaining shifts in the demand function mightbe expected to consider industrial production as such or the ac-tivity of an individual steel-consuming industry. But this ap-proach can be refined. Our method refines it by taking into ac-count not only the over-all level of production, but also itscyclical position. The cyclical variation in the "product-mix" ofindustrial output and variation in the "working-inventory re-quirements" will be reflected in the short-term component. Thisrefinement of the approach leads to Formula II, which providesnot only an improved explanation of the demand for steel duringthe period of observation, 1919-40, but also leads to radicallydifferent and far more accurate results when used for long-runextrapolation. A further attempt at estimating the effect of tech-nological progress in steel consumption within the 22 years underconsideration resulted in Formula III.
There was an indication that the output destined for the for-eign market was often distorted substantially by exogenous fac-tors not related to the level of domestic industrial activity. Sinceforeign demand accounted for a fluctuating share of the totalsteel output, varying from 8 percent to 18 percent, the outputfor the domestic market was investigated separately. This re-lation is described by Formula IV.
256
STEEL REQUIREMENTS
Finally, it seemed useful to explore the relation of steel ca-pacity to over-all industrial capacity, both to gauge the adequacyof steel capacity during the period of observation as well asthereafter and to gain a better understanding of the concept andstatistical construction of industrial capacity. Formula V expressesindustrial peacetime capacity in terms of steel requirements. Aninversion of the process could be used to determine the cyclicalexpansion permitted by the level of rated steel capacity, but thisis subject to the qualification that the estimate of steel require-ments per unit of output is based on peacetime goods. If indus-trial production comprises a large portion of combat materialother than ships, unit steel requirements are lowered and allowa considerable extension of the area of cyclical expansion. Thishappened during World War II when more intensive use of fa-cilities also caused industrial activity to exceed the peacetimelevel of industrial capacity.
The successive steps of this analysis are: (1) distinguishingbetween the long-term and short-term components of demand,(2) distinguishing between conditions of steel demand in thetwenties and the thirties, and (3) distinguishing between exportand domestic demand. Some of the procedures as well as sometools of the analysis are of a tentative character and subject tofurther improvements, but it is hoped that the general outline ofthe procedure will be useful for a similar analysis of raw ma-terials or finished products.
APPENDIX APeacetime Civilian Capacitq for industrial Production
The series on peacetime civilian capacity for industrial produc-tion used in the analysis of steel requirements is a revision of anearlier series, developed by C. F. Roos, V. V. Szeliski, and F. L.Alt before World War JJ•12 A more detailed account of the con-ceptual problems, construction, and limitations of this series willbe found in a forthcoming paper. The description presented hereis confined to the essential elements.
12 The series forms an integral part of the copyrighted service of theEconometric Institute, Inc. The data and the series are presented here bypermission of C. F. Roos, president of the Institute.
The basic principles of the capacity series were used by Boos in somewhatcruder form in 1937 in order to point out that in spite of an army of nearly8 million unemployed, the mechanical facilities were becoming scarce andthat, therefore, demand for investment goods would take a sharp upturn.
257
STEEL REQUIREMENTS.
The series is constructed on the basis of annual data on installa-tion of equipment in manufacturing and mining given in 1935-39prices (column 1, Table A-i). It is essentially a portion of theproducers' equipment data of the gross national product accountof the Department of Commerce's deflated by H. Shavell's priceindexes for capital equipment.'4 The series has been carried backwith the help of data developed by S. Kuznets and W. H. Shaw.15Data after 1945 have been derived from subsidiary informationand are only tentative.
The efficiency factor shown in column 2 of Table A-i convertsthe dollar volume of equipment installed (column 1) into termsof gross addition to capacity .( column 8). A gradual linear in-crease in initial efficiency has been assumed. This increase isimplicitly given once the bench marks for capacity are set, andthe pattern and length of retirement assumed. The retirementpattern applied to gross addition to capacity assumes a full lifeof 16 years before 1939 and 14 years for the later period. It is acompromise between a straight-line retirement for 16 years andan 8-year step function dropping from 100 to 0 percent survival.A one-parametric sequence of such retirement has been de-veloped by B. F. Kimball.16 The first eight years of the symmetricretirement pattern for 16 years indicate the following survivalvalues: 1.000, 1.000, 0.994, 0.977, 0.983, 0.841, 0.692, 0.500. Column4 of Table A-i represents the portion of capacity retired duringcalendar year t. If column 4 is subtracted from the gross ad-dition for the same year (column 3), the net addition to capacity(column 5) is derived. Going forward and backward from thebench mark level of 126 for the end of 1929, the annual capacityseries is computed (column 6). Interpolating the midyear ca-pacity, together with total industrial production for the calendaryear (column 8), determines the rate of capacity use (column 9).
13 "National Income Supplements" to the Survey of Current Business,July 1947, table 32, p. 45; and July 1950, p. 26.
14 Henry Shavel!, "Price Deflators for Consumer Commodities and CapitalEquipment, 1929-1942," Survey of Current Business, May 1943, pp. 13ff.The deflation of producers' equipment by segments was undertaken beforethe appearance of "Estimates of Gross National Product in Constant Dollars,1929-1949," Survey of Current Business, January 1951, p. 9.
15 Simon Kuznets, Commodity Flow and Capital Formation, Vol. I(NBER, 1938); William H. Shaw, Value of Commodity Output since 1869(NBER, 1947).
Bradford F. Kimball, "A System of Life Tables for Physical Property,"Econometrica, September 1947.
258
—
STEEL REQUIREMENTS
The capacity series presented here is an over-all series. It is notan aggregation of capacities in individual industries, which, bynecessity, must be higher than the over-all series so long as im-balances between consuming and supplying industries exist. Itsbench marks have been developed for years of peak production.Assuming that "normal" production is characterized by 15 percentunused, stand-by capacity to take care of seasonal and other peaksduring the year or to meet emergencies, the "peak" rate of opera-tions was set halfway between the 85 percent "normal" rate ofoperations and the 100 percent capacity rate of operations.
Paralleling the analysis of steel demand is an estimate of therelative importance of equipment for manufacturing and miningin total industrial production, based on the rate of operations dur-ing the preceding fiscal year. Introducing PEMM for producers'equipment for manufacturing and mining in points of the totalproduction index, and
171the rate of operations in percent of capacity, we haveFormula VI
PEMM/RT = = 1.794 + 0.479 •
The volume of equipment for manufacturing and mining istranslated from constant 1985-39 dollars into terms of the pro-duction index by a conversion factor of $315.8 million (1935-39dollars) per point on the index developed for the output of capitalgoods as a whole. Table A-2 contains the necessary annual dataand indicates the computations. Except for 1937, the war years1944-45, when investment anticipated reconversion, and 1950, theyear in which defense investment was accentuated by fears ofmaterial shortages, the actual observations differ from the cal-culated relative importance of producers' equipment by not morethan 0.7 percent of the production index. These limits are shownin Chart A-i, which compares the output of equipment for manu-facturing and mining industries (in percent of total industrialproduction) with the rate of operations during the precedingfiscal year (in percent of capacity).
This is rather remarkable, since the dollar volume of newequipment installed is stated in nondescript constant dollars: itdoes not say whether $100 worth of equipment is for a steel millor a candy factory. It cannot be expected that the output po-
259
TAB
LE A
-iC
APA
CIT
Y JN
MA
NU
FAC
TUR
ING
AN
D M
ININ
G, 1
918-
1950
Equi
pmen
tIn
stal
led
(inM
illio
ns o
fD
olla
rs, 1
935-
Effic
ienc
y,(t
—19
00)
Gro
ssC
apac
ityN
etC
apac
ityC
apac
ityat
End
Add
edof
Yea
r
Cap
acity
0$ o
fJu
ly 1
FRB
inde
xof
Pro
duc-
tion
(193
5-
Rat
e of
Ope
ratio
n,(8
) ÷(7
)A
dded
Ret
ired
39 Prices)
x 0.
1707
87(INPOINTS OFTHE FRB INDEX, 1935-39
=100)
39
100)
(in P
erce
nt)
Yea
r(1
)(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
1918
1,59
73.
074
4.91
1.49
3.42
88.3
419
191,
398
3.24
54.
531.
682.
8691
.20
89.7
772
80.2
1920
1,54
13.
418
5.28
1.88
3.38
94.5
892
.89
7580
.719
2188
93.
587
3.19
2.11
1.08
95.6
695
.12
5861
.019
221,
259
3.75
74.
732.
412.
3297
.98
96.8
273
75.4
1923
1,67
53.
928
6.58
2.80
3.78
101.
7699
.87
8888
.119
241,
468
4.09
96.
023.
282.
7410
4.50
103.
1382
79.5
1925
1,69
74.
270
7.25
3.77
3.48
107.
9810
6.24
9084
.719
261,
827
4.44
08.
114.
183.
9311
1.91
109.
9596
87.3
1927
1,67
24.
611
7.71
4.45
3.26
115.
1711
3.54
9583
.719
281,
892
4.78
29.
054.
614.
4411
9.61
117.
3999
84.3
1929
2,25
64.
953
11.1
74.
788.
3912
6.00
122.
8011
089
.619
301,
647
5.12
48.
445.
083.
3612
9.36
127.
6891
71.3
1931
1,18
15.
294
8.25
5.55
0.70
130.
0612
9.71
7557
.819
3277
25.
465
4.22
6.14
—1.
9212
8.14
129.
1058
44.9
1933
755
5.63
64.
266.
77—
2.51
125.
6312
8.88
6954
.419
34967
5.807
5.61
7.39
—1.77
123.86
124.75
7560.1
1935
1,30
75.
978
7.81
7.94
—0.
1312
3.73
123.
7987
70.3
1936
8.14
811
.72
8.32
3.40
127.
1312
5.43
108
82.1
1937
1,91
86.
319
12.1
28.
363.
7613
0.89
129.
0111
387
.619
381,
407
6.49
09.
137.
991.
1413
2.03
131.
4689
67.7
1939
1,63
56.
661
10.8
97.
313.
5813
5.61
133.
8210
981
.5
TAB
LE A
-i (c
oncl
uded
)
U, til tTI
tTi
40 tTj z '-3 U)
I.
Equi
pmen
tin
stal
led
(inM
illio
ns o
fD
olla
rs, 1
935-
Effic
ienc
y,(t
— 1
900)
Gro
ssC
apac
ityN
etC
apac
ityC
apac
ityat
End
Add
edof
Yea
r
Cap
acity
as o
fJu
ly 1
FRB
Inde
xof
Pro
duc-
tion
(193
5-
Rat
e of
Ope
ratio
n,(8
) ÷ (7
)A
dded
Ret
ired
39 P
rices
)X
0.1
7078
7(I
NPO
INTS
OF
THE
FRB
IND
EX, 1
935-
39=
100)
39 =
100)
(in P
erce
nt)
Yea
r(1
)(2
)(3
)(4
)(5
)(6
)(7
)(8
)(9
)19
402,
146
6.83
114
.66
7.48
7.18
142.
7913
9.20
125
89.8
1941
2,56
07.
002
17.9
37.
4810
.45
153.
2514
8.02
182
109.
519
421,
586
7.17
311
.88
7.98
3.40
156.
8515
4.95
199
.12
8.4
1943
1,27
87,
343
9.38
9.35
0.03
156.
6815
6.67
239
152.
519
441,
794
7.51
513
.48
10.9
92.
4915
9.17
157.
9323
514
8.8
1945
2,43
67.
685
18.7
211
.77
6.95
166.
1216
2.65
203
124.
819
463,
400
7.85
826
.71
12.6
814
.03
180.
1517
8.14
170
98.2
1947
3,80
08.
027
30.5
013
.11
17.3
919
7.54
188.
8518
799
.019
483,
700
8.19
830
.33
13.4
516
.88
214.
4220
5.98
192
93.2
1949
3,50
08.
369
29.2
918
.62
15.6
728
0.09
222.
2617
879
.219
504,
150
8.53
935
.43
16.0
819
.35
249.
4523
8.6
200
83.8
TAB
LE A
-2R
ECU
IISI
ON
FO
R T
HE
REN
EWA
L A
N])
GR
owm
OF
PRO
DU
CTI
ON
CA
PAC
ITY
, 191
9-19
50
Shar
e of
Equ
ipm
ent f
orM
anuf
actu
ring
and
Min
ing
inC
ivili
an In
dust
rial P
rodu
ctio
n(P
erce
nt)
Peac
etim
eTo
tal
Civ
ilian
Out
put o
f Equ
ipm
ent f
orC
ivili
anC
ivili
anR
ate
ofC
alcu
late
d,In
dust
rial
Man
ufac
turin
g an
d M
inin
g in
Poi
nts o
f the
Cap
acity
Indu
stria
lO
pera
tions
St (
rjj)
Prod
uctio
nIn
dex
of T
otal
indu
stria
l Pro
duct
ion
(Beg
inni
ng o
fPr
oduc
tion
(Fis
cal Y
ear,
= 1.
794
+A
ctua
l,(C
alen
dar
Cal
enda
r Yea
r), (
Fisc
al Y
ear)
Perc
ent),
0.04
798*
Yea
r)C
alcu
late
d,A
ctua
l,D
evia
tion,
PCA
P(t —
1)PT
(t —
(8)
(6)
PT(t)
(4) X
(6)
PEM
M(t)
(8) —
(7)
Yea
r(1
)(2
)(3
)(4
)(5
)(6
)(7
)(8
)(9
)
1919
7180
.45.
656.
1572
4.07
4.43
+0.3
619
2091
.277
84.4
5.84
6.51
754.
384.
88+0
.50
1921
94.6
6467
.75.
044.
8558
2.92
2.81
—0.
1119
2295
.764
66.9
5.00
5.47
733.
653.
99+0
.84
1923
98.0
8485
.75.
906.
0288
5.19
5.30
+0.1
119
2410
1.8
8583
.55.
795.
6782
4.75
4.65
—0.
10
1925
104.
585
81.3
5.69
5.96
905.
125.
36+0
.24
1926
108.
098
86.1
5.92
6.03
965.
685.
79+0
.11
1927
111.
997
86.7
5.95
5.56
955.
655.
29—
0.36
1928
115.
295
82.5
5.75
6.09
995.
695.
99+0
.80
1929
119.
610
789
.56.
086.
4911
06.
697.
14+0
.45
1930
126.
010
381
.75.
715.
7391
5.20
5.22
+0.0
219
3112
9.4
8263
.44.
834.
9875
3.62
3.74
+0.1
219
3218
0.1
6550
.04.
194.
2158
2.43
2.44
+0.0
119
3312
8.1
6046
.84.
043.
4669
279
2.89
—0.
4019
3412
5.6
7660
.54.
694.
0875
3.52
3.06
—0.
48
1935
123.
978
63.0
4.81
4.75
874.
184.
14—
0.04
1936
123.
794
78.0
5.43
5.85
103
5.59
6.03
1937
127.
111
489
.76.
095.
3611
36.
886.
07—
0.81
1938
130.
995
72.6
5.27
502
894.
694.
48—
0.23
1939
a18
2.0
9874
.25.
354.
8010
85.
785.
18—
0.60
TAB
LEA
-2. (
conc
lude
d)Sh
are
of E
quip
men
t for
Man
ufac
turin
g an
d M
inin
g in
Civ
ilian
indu
stria
l Pro
duct
ion
(Per
cent
).
Peac
etim
eTo
tal
Civ
ilian
Civ
ilian
Cap
acity
indu
stria
l(B
egin
ning
of
Prod
uctio
nC
alen
dar Y
ear)
, (Fi
scal
Yea
r)
Rat
e of
Ope
ratio
ns(F
isca
l Yea
r,Pe
rcen
t),
Civ
ilian
Indu
stria
lPr
oduc
tion
(Cal
enda
rY
ear)
Out
put o
f Equ
ipm
ent f
orM
anuf
actu
ring
and
Min
ing
in P
oint
s of t
heIn
dex
of T
otal
Indu
stria
l Pro
duct
ion
Cal
cula
ted,
s, ( 1.
794
+A
ctua
l,0.
0479
s'C
alcu
late
d,A
ctua
l,D
evia
tion,
Yea
rPC
AP(
t —1)
—1)
(1)
(2)
(3)
(8) ÷
(6)
(4)
(5)
PT(t) (6)
(4) X
(6)
PBM
M(t)
(8) —
(7)
(7)
(8)
(9)
1940
"13
,5.6
113.
583
.75.
805.
7111
96.
906.
79—
0.11
1941
a14
2.8
124.
587
.25.
976.
2313
07.
768.
10+0
.34
1942
"15
3.2
111
72.5
5.27
5.48
924.
855.
03+0
.18
1943
"15
6.7
8654
.94.
425.
0680
3.54
4.05
+0.5
1J9
44a
156.
780
51.1
4.24
7.10
803.
395.
68+2
.29
1945
a15
9.2
8150
.94.
237.
8990
3.81
7.71
+3.9
019
4616
6.1
137
82.5
5.75
6.34
170
9.78
10.7
8+0
.98
1947
180.
218
310
1.6
6.66
6.44
187
12.4
512
.04
—0.
4119
4819
7.5
189
95.7
6.38
6.10
192
12.2
511
.71
—0.
5419
4921
4.4
188
86.8
5.95
6.35
175
10.4
111
.09
+0.6
819
5023
0.1
180
78.2
5.54
8.80
200
11.0
813
.20
+2.1
2-
a C
ivili
anpr
oduc
tion
as in
dica
ted
in ta
ble
on p
age
852
of th
e Fe
dera
l Res
erve
Bul
letin
, Sep
tem
ber
1945
, for
the
cale
ndar
yea
rs 1
939
thro
ugh
1945
.
STEEL REQUIREMENTS
CHART A-iOutput of Manufacturing and Mining Equipment Compared
with Rote of Operation of Total Industrial Capacity
Equipment output as percentof totat production
Output of equipment for manufacturing and mining is measured as a share oftotal industrial production during the calendar year.
Rate of operotion is measured as the portion of peacetime civilian capacity usedby total industrial production during the preceding fiscal year.
Estimates for the years 1939-45 cover only nonwar industrial production and areindicated by the dashed line (figures from the Federal Reserve Bulletin, September1945).
The Formula VI estimate of manufacturing and mining equipment output comparedwith rate of operation is indicated by the heavy line. Note that in most years therange of variation of actual data from this estimate is not more than ±0.7 of 1 per-cent (indicated by the dot-dashed lines).
tential will be the same in each case. The same holds for capacityitself. The set of conditions accompanying the capacity levels inbench mark years is not necessarily duplicated in the unknowncomposition of annual gross addition to capacity. Thus, develop-
264
60 65 70 90Rate of operatLon as percent of totql LndustrLal capqcLty
STEEL REQUIREMENTS
ing capacity levels only on the basis of a nondescript constant-dollar volume of installed new equipment may lead at times toan unattainable level of capacity. This was demonstrated by theconstraint imposed by steel capacity shown in Formula V, in Sec-tion H. However, it must be assumed that such deviations aremore or less short-term in character. Thus, given a long-term pro-jection of industrial production, Formula VI permits a recursivedevelopment of a long-term projection of capacity. This provides,in turn, a partial test of the long-term projection of industrial pro-duction itself, since the changes in the internal structure of in-dustrial production during the business cycle are of a fafrlydefinite character. They will permit a check of the estimates ofproducers' equipment for consistency.
265
APP
END
IX B
TAB
LE B
-iST
EEL
PRO
DU
CTI
ON
, CA
PAC
ITY
, AN
D R
EQU
IREM
ENT
RA
rES
IN T
HE
UN
ITED
STA
TES,
191
9-19
51
PRO
DU
CTI
ON
OF
ING
OTS
AN
D S
TEEL
FO
RPR
OD
UC
TIO
N C
APA
CIT
YC
AST
ING
S D
ESTI
NED
FO
R V
AR
IOU
S M
AR
KET
SFO
R IN
CO
TS A
ND
STE
ELA
PPA
REN
T B
ATE
OF
(MIL
LIO
NS
OF
NET
TO
NS)
FOR
CA
STIN
GS,
MID
YEA
RST
EEL
REQ
umEM
ENTS
aTo
tal,
Dom
estic
,Fo
reig
n,S.
f.SD
(MIL
LIO
NS
OF
NET
TO
NS)
SOA
PTo
tal,
Dom
estic
,ST
/PT
SD/P
TY
EAR
(1)
(2)
(3)
(4)
(5)
(6)
1919
38.8
31.7
7.1
61.7
539
440
1920
1921
1922
47.2
22.2
39.9
39.7
19.0
36.8
7.5
3.2
3.1
83.3
64.9
65.6
629
383
547
529
328
504
1923
1924
1925
1926
50.3
42.5
50.8
54.1
47.1
39.8
48.0
50.7
3.2
2.7
2.8
3.4
66.2
67.6
66.7
66.0
572
518
564
564
535
485
533
528
1927
1928
50.3
57.7
47.1
53.9
3.2
3.8
68.0
70.1
529
583
496
544
1929
63.2
59.2
4.0
72.7
575
538
1930
45.6
43.0
2.6
75,2
501
473
1931
29.1
27.7
1.4
78.1
888
369
1932
15.3
14.7
.678
.726
425
319
3326
.025
.1.
.978
.437
736
419
34.
29.2
27.6
1.6
78.3
389
368
1935
38.2
36.6
1.6
78.3
439
421
1936
53.5
51.1
2.4
78.4
519
498
1937
56.6
50.8
5.8
79.4
501
450
1938
31.8
28.4
8.4
81.0
357
319
1939
52.8
48.1
4.7
81.7
484
441
TAB
LEB
-i (c
oncl
uded
)
YEA
R
PRO
DU
CTI
ON
OF
ING
OTS
AN
D S
TEEL
FO
RC
AST
ING
S D
ESTI
NED
FO
R V
AR
IOU
S M
AB
EETS
OF
NET
TO
NS)
Tota
l,D
omes
tic,
Fore
ign,
STSE
(1)
(2)
(3)
PRO
DU
CTI
ON
CA
PAC
ITY
FOR
ING
OTS
AN
D S
TEEL
FOR
CA
STIN
GS,
MID
YEA
R(M
ILLI
ON
S O
F N
ET T
ON
S)
(4)
APP
AR
ENT
BA
TE O
FST
EEL
REQ
uIR
EMEN
Tsa
Tota
l,D
omes
tic,
ST/P
TSD
/PT
(5)
(6)
1940
1941
1942
1943
1944
67.0
54.8
12.2
82.8
73.4
9.4
86.0
75.7
10.3
88.8
78.3
10.5
89.6
81.3
8.3
83.4
87.0
89.4
91.9
94.7
536
438
511
453
432
380
372
328
381
346
1945
1946
1947
1948
1949
79.7
72.4
7.3
66.6
59.7
6.9
84.9
75.9
9.0
88.6
82.2
6.4
78.0
70.6
7.4
93.7
91.6
92.7
95.2
97.8
893
357
392
851
454
406
461
428
443
401
1950
1951
96.8
92.7
4.1
105.
110
0.5
4.6
100.
648
446
347
845
7
a In
thou
sand
s of n
et to
ns a
nnua
lly p
er p
oint
on
the
Fede
ral R
eser
ve In
dex
of In
dust
rial P
rodu
ctio
n.
Col
umn
1So
urce
: Am
eric
an Ir
on a
nd S
teel
Inst
itute
, Met
al S
tatis
tics,
1950
, p. 1
19.
2(S
1—S5
)S
This
is a
n es
timat
e ba
sed
on th
e pe
rcen
tage
whi
ch "
Iron
and
Ste
el E
xpor
tsO
ther
than
Scra
p" (D
epar
tmen
t of C
omm
erce
, Met
al S
tatis
tics,
1950
, p. 2
60)
isof
"R
olle
d St
eel
Prod
ucts
for S
ale"
(Am
eric
an Ir
on a
nd S
teel
Inst
itute
, op.
cit.,
p. 1
35) a
pplie
d to
col
umn
1.B
oth
serie
s are
stat
ed fo
r the
12-
mon
th p
erio
d en
ding
Janu
ary
31. T
he c
ompa
rison
ofr
olle
dst
eel f
or sa
le w
ith p
rodu
ctio
n of
ingo
ts a
nd st
eel f
or c
astin
gs im
plic
it in
this
pro
cedu
reac
coun
ts (a
) for
the
rate
of s
crap
and
(b) f
or th
e sh
are
of ro
lled
stee
l not
for s
ale
(but
for r
econ
vers
ion)
. Bot
h ra
tes v
ary
over
tim
e.3.
74
Rep
orte
d or
est
imat
ed fo
r mid
year
as a
n ap
prox
imat
ion
for t
he a
nnua
l ave
rage
. Sou
rce:
Sam
e as
col
umn
1.72
.95
Col
umn
1 ÷
colu
mn
1 of
Tab
le A
-2.
6C
olum
n 2
÷co
lum
n1
of T
able
A-2
.
to'C
Ave
rage
,19
19-4
043
.740
.0
TAB
LE B
-2TO
TAL
IND
UST
BIA
L A
ND
SEL
ECTE
D M
ETA
L PR
OD
UC
TIO
N, 1
919-
1951
(193
5-39
—10
0)
Prod
uctio
nIn
dust
rial P
rodu
ctio
nM
etal
and
Iron
and
Peac
etim
eTo
tal
Indu
stria
l,M
etal
Prod
ucts
,St
eel
Prod
ucts
,N
orm
al L
evel
,O
.85(
PO
AP)
Cyc
lical
Slac
k,C
ivili
an C
apac
ity(M
idye
ar),
PTPM
ETor
PN
POT
PCA
PY
ear
(1)
(2)
(3)
(4)
(5)
(6)
1919
7224
.29.
277
590
1920
1921
1922
1923
1924
75 58 73 88 82
27.3
13.6
22.9
30.2
27.2
11.2 5.3
9.4
12.0 9.9
79 81 82 85 88
4 23 9—
3 6
93 95 97 100
103
1925
1926
1927
1928
1929
90 96 95 99 110
31.1
33.6
31.1
34.9
40.4
11.9
12.7
11.9
13.3
14.6
90 94 97 99 105
0—
2 2 0—
5
108
110
114
117
123
1930
1931
1932
1933
1934
91 75 58 89 75
29.7
19.8
11.8
16.1
19.9
10.7 6.7
3.5
5.9
6.7
109
111
110
108
106
18 36 52 39 31
128
130
129
127
125
I
TAB
LE B
-2 (c
oncl
uded
)
tIJ tTl
,0 c tTj z C
l)
Prod
uctio
nIn
dust
rial P
rodu
ctio
nM
etal
and
iron
and
Peac
etim
eTo
tal
indu
stria
l,M
etal
Prod
ucts
,St
eel
Prod
ucts
,N
orm
al L
evel
,0.
85(P
CA
P)C
yclic
alSl
ack,
Civ
ilian
Cap
acity
(Mid
year
),PT
PMET
or P
NPO
TPC
AP
Yea
r(1
)(2
)(3
)(4
)(5
)(6
)19
3587
25.4
8.9
105
1812
41936
103
33.3
12.6
106
3125
1937
113
379
13.5
110
—3
129
1938
89
22.9
7.5
111
22
131
1939
109
33.1
12.5
114
5134
1940
125
43.8
16.2
118
—7
139
1941
162
65.5
20.5
126
—36
148
1942
199
94.8
21.9
132
—67
155
1943
239
125.6
22.9
133
—106
157
1944
235
123.5
22.7
134
—101
158
1945
203
94.2
20.1
139
—64
163
1946
170
61.4
18.6
147
—23
173
1947
187
70.9
21.5
161
—26
189
1948
192
72.8
22.9
175
—17
206
1949
176
64.9
20.6
•189
13
222
1950
200
77.0
24.8
203
323
8.6
1951
220
88.8
28.5
222
2261.7
Col
umn
1So
urce
: Boa
rd o
f Gov
erno
rs o
f the
Fed
eral
Res
erve
Sys
tem
, Fed
eral
Res
erve
Bul
letin
, Oct
ober
194
3.2
Sour
ce: S
ame
as c
olum
n 1.
3So
urce
: Sam
e as
col
umn
1.4
Sour
ce: S
ee A
ppen
dix
A.
5B
ecau
se o
f the
pre
vale
nce
of sl
ack
rath
er th
an st
rain
(—P
cr=
PT
—P
N)
durin
g th
e pe
riod
1919
-40,
cycl
ical
slac
k ha
s bee
n us
ed a
s the
shor
t-ter
m v
aria
ble.
6So
urce
: See
App
endi
x A
.
Ave
rage
191
9-40
87.8
27.7
10.3
99.3
11.5
116.
8
TAB
LE B
-3 OF
TOTA
L ST
EEL
PRO
DU
CTI
ON
AN
D O
F PR
OD
UC
TIO
N O
F ST
EEL
FOR
TH
ED
OM
ES
TIC
MA
RK
ETB
ASE
D O
N IK
DtJ
STB
IAL
PRO
DU
CTI
ON
AN
!) IT
s NO
RM
AL
1919
-193
9 (iw
MIL
LIO
NS
OF
TON
S)
Form
ula
IFo
rmul
a ii
Form
ula
III
Form
ula
JV
Year
ST(PT)
(1)
Res
idua
ls,
ST—
(1)
(2)
ST(PT,PN)
(3)
Res
idua
ls,
ST —
(3)
(4)
ST(5
)(6
)SD
(PT,
PN)
(7)
SE + SD(PP,PN)
(8)
Res
idua
ls,
ST —
(8)
(9)
1919
321
-L8
39.7
-0.9
356
42.7
-3.9
1920
84.8
+ 12
.442
.841
.7+5
.537
.745
.2+2
.019
2122
.9—
0.7
26.7
—4.
5—
1.7
18.5
21.7
+0.5
1922
83.4
+6.5
39.1
+0.8
38.3
+1.6
34.0
37.1
+2.8
1923
43.8
+6.5
50.6
—0.
351
.7—
1.4
48.4
51.6
—0.
719
24
1925
40.0
45.2
4-2.
5+5
.644
.049
.9—
1.5
+0.9
44.3
51.2
—1.
8—
0.4
40.3
47.8
43.0
50.8
—0.
5+0
.219
2649
.4+4
.753
.1+1
.055
.2—
1.1
52.0
55.4
—L3
1927
48.7
+1.6
50.9
—0.
652
.8—
2.5
49.3
52.5
—2.
219
2851
.5+6
.253
.355
.8+1
.952
.556
.3+1
.419
2959
.2+4
.059
.9+3
.361
.9+1
.357
.861
.8+1
.419
3045
.9—
0.3
41.7
4-3.
941
.74-
8.9
37.8
40.4
+5.2
1931
34.8
—5.
727
.0+2
.127
.7+1
.426
.928
.3+0
.819
3222
.9—
7.8
•12
.9•
+2.4
14.5
+0.8
16.2
16.8
—1.
519
3330
.6—
4.6
23.4
+2.6
24.0
+2.0
23.6
24.5
+1.5
1934
34.8
—5.
629
.4—
0.2
29.4
—0.
227
.929
.5—
0.3
1935
43.1
—4.
940
.2—
2.0
39.3
—1.
185
.737
.3+0
.919
3854
.3—
0.8
53.4
+0.1
51.6
+1.9
45.8
48.2
+5.3
1937
61.3
—4.
760
.1—
3.5
58.2
—1.
651
.457
.2—
0.6
1938
44.5
—12
.739
.0—
7.2
88.8
—7.
035
.839
.2—
7.4
1939
58.5
—5.
754
.8—
2.0
58.7
—0.
948
.052
.7+0
.1
Cl) -3 r -4 tTJ '-3 C
l)
TAB
LE B
-3 (c
oncl
uded
)
U)
'-3
LTI
tTj I
I'3
Form
ula
IFo
rmul
aII
Form
ula
III
Form
ula
W
Res
idua
ls,
Res
idua
ls,
Res
idua
ls,
Res
idua
ls,
ST(PT)
ST —
(1)
ST(PT,PN)
ST —
(3)
ST(PT,PN)
ST —(5)
SD(PT,PN)
SE +
Sn(P
T,PN
)ST —
(8)
Yea
r(1
)(2
)(3)
(4)
(5)
(6)
(7)
(8)
(9)
1940
69.6
—2.
666
.8+0
.465
.0+2
.057.5
69.7
—2.
719
4193
.4—
12.6
94.6
—11
.891
.7—
8.9
79.5
88.9
—6.
119
4212
1.2
—35
.212
3.4
—37
.411
9.1
—33
.110
2.1
112.
4—
28.4
1943
149.
1—
60.3
157.
1—
68.3
.15
0.5
—61
.612
7.4
137.
9—
49.0
1944
146.
3—
56.7
153.
3—
63.7
147.
0—
57.4
124.
713
3.0
—43
.419
4512
4.0
—44
.312
3.4
—43
.711
9.9
—40
.210
3.2
110.
5—
30.8
1946
101.
0—
34.4
91.4
—24
.891
.0—
24.4
80.6
87.5
—20
.919
4711
2.8
—27
.999
.4—
14.5
99.8
—14
.988
.797
.7—
12.8
1948
118.
3—
27.7
97.0
—8.
499
.0—
10.4
89.2
95.6
—7.
019
4910
5.2
—27
.276
.7+1
.381
.6—
3.7
76.2
83.6
—5.
7
1950
121.
9—
25.2
90.8
+6.2
95.9
+0.9
88.7
92.8
+4.0
1951
135.
9—
30.8
98.7
+6.4
104.
6+0
.597
.810
2.4
+2.7
STEEL REQUIREMENTS
TABLE B-4ESTIMATES OF STEEL REQUIREMENTS AT NORMAL LONG-RUN CAPACITYRATES OF TOTAL INDUSTRIAL PRODUCTION BASED ON VAiuous FORMULAE,1919-1951 (IN MILLIONS OF TONS)
NORMAL LONG-RUN RATES CAPACITY RATESBASEI) ON FORMULA BASE]) ON FORMULA
ii III 1V' II IIIST(PN) ST(PM) SD(PN) ST(POAP) ST(PCAP)
YEAR (1) (2) (3) (4) (5)
1919 44.9 44.6 41.0 56.3 57.8
1920 45.7 45.7 42.0 57.6 59.5
1921 46.4 46.7 43.1 58.5 60.7
1922 46.8 47.2 43.6 59.4 61.8
1923 48.0 48.7 45.2 60,8 63.6
1924 49.1 50.2 46.7 62.2 65.3
1925 49.9 51.2 47.8 63.5 67.0
1926 51.4 53.3 49.9 65.8 69.3
1927 52.6 54.8 51.4 67.2 71.7
1928 53.3 55.8 52.5 68.5 73.4
1929 55.6 57.3 53.1 71.2 74.2
1930 57.1 57.0 51.6 73.5 73.3
1931 57.9 56.3 49.9 74.4 71.5
1932 57.5 55.8 49.5 74.0 71.0
1933 56.8 54.9 48.6 73.1 70.0
1934 56.0 54.0 47.7 72.2 69.0
1935 55.6 53.5 47.2 71.7 68.5
1936 56.0 54.0 47.7 72.2 69.0
1937 57.5 55.8 49.5 74.0 71.0
1938 57.9 56.3 49.9 74.9 72.0
1939 59.1 57.7 51.2 76.2 73.6
1940 60.6 59.5 53.0 78.5 76.1
1941 63.7 63.1 58.5 82.6 80.7
1942 66.0 65.9 59.2 85.8 84.2
1948 66.3 66.3 59.6 86.7 85.2
1944 66.7 86.8 60.1 87.1 85.7
1945 68.6 69.1 62.8 89.4 88.3
1946 71.7 72.7 65.9 93.9 93.4
1947 77.1 79.1 72.1 101.2 101.5
1948 82.4 85.5 78.3 108.9 110.1
1949 87.8 91.9 84.5 118.2 118.2
1950 93.1 98.3 90.7 123.7 128.6
1951 100.4 106.9 99.1 134.5 138.5
272