This PDF is a selection from an out-of-print volume from the NationalBureau of Economic Research

Volume Title: The Structure of Manufacturing Production: A Cross-SectionView

Volume Author/Editor: Charles A. Bliss

Volume Publisher: UMI

Volume ISBN: 0-87014-035-3

Volume URL: http://www.nber.org/books/blis39-1

Publication Date: 1939

Chapter Title: IV Summary: The Pattern of Manufacturing Production

Chapter Author: Charles A. Bliss

Chapter URL: http://www.nber.org/chapters/c6996

Chapter pages in book: (p. 120 - 140)

IvSummary: The Pattern of Manufacturing

Production

IN 1791 Alexander Hamilton presented to the new Congresshis famous report on Manufactures. In doing so, he listedsome seventeen groups of manufacturir1g industries which,in his words, had CCgrown up and flourished with a rapiditywhich surprises, affording an encouraging assurance of suc-cess in future attempts".' But even Hamilton could not fore-see the success that did attend these efforts, or the transitionby which the industries of 1791, whose products ranged from1The 17 commodity groups are: "i) Of Skins—Tanned and tawed leather, dressedskins, shoes, boots and slippers, harness and saddlery of all kinds, portmanteaus andtrunks, leather breeches, gloves, muffs and tippets, parchment and glue. 2) 0/ Iron—Bar and sheet iron, steel, nail rods and nails, implements of husbandry, stoves, pots,and other household utensils, the steel and iron work of carriages, and for ship-building, anchors, scale beams and weights, and various tools of artificers, arms ofdifferent kinds; though the manufacture of these last has of late diminished for wantof demand. 3) Of Wood—Ships, cabinet wares, and turnery, wool and cotton cards,and other machinery for manufactures and husbandry, mathematical instruments,coopers' wares of every kind. 4) Of Flax and Hemp—Cables, sail-cloth, cordage,twine, and pack thread. Bricks and coarse tiles, and wares. 6) Ardentspirits, and malt liquors. 7) Writing and printing paper, sheathing and wrappingpaper, paste boards, fullers' or press papers, paper-hangings. 8) Hats of fur and wool,and mixtures of both; women's stuff and silk shoes. 9) Refined sugars. io) Oils ofanimals and seeds, soap, spermaceti and tallow candles. ix) Copper and brass wares,particularly utensils for distillers, sugar refiners and brewers ; andirons and otherarticles for household use, philosophical apparatus. iz) Tinwares for most purposesof ordinary use. 13) Carriages of all kinds. 14) Snuff, chewing and smoking tobacco.

Starch and hair powder. x6). Lamp-black and other painters' colors. 17) Gun-powder." In addition, the report stated, there was a "vast scene of household manu-facturing, which contributes more largely to the supply of the community than couldbe imagined without having made it an object of particular inquiry".

Hamilton's report and his correspondence on the subject of manufactures have beenpublished under the editorship of A. H. Cole in Industrial and Commercial Corre-spondence of Alexander Hamilton (Business Historical Studies, Vol. I, A. W. ShawCo., 1928). (Citations are from pp. 279—80.)

SUMMARY 121

starch and hair powder to various articles of iron, became themanufacturing system of 1929. During the intervening onehundred thirty-eight years the scale of. operations, thetechnology of manufacturing processes, the markets served,have all been profoundly modified, the flow of goods tre-mendously increased. New areas of manufacturing activityhave been developed, new products undreamed of—theautomobile, the various forms of electrical apparatus,petroleum and its distillates, rayon. The making of lamp-black, so encouraging to Hamilton at the turn of theeighteenth century, is today completely overshadowed by themaking of carbon black (from gas instead of coal or woodtar) produced to meet the needs of the automobUe tireindustry. Virtually all the articles of household manufacturein 1791 are now made in factories; more and more house—hold duties have been supplanted by services originating inthe manufacturing field or have been made less burdensomeby reason of investment in factory-made equipment. Inevery way the pattern of manufacturing production has be-come more elaborate.

Just over a hundred years after Hamilton's report ap-peared the comprehensive Census of i 899, from which thepresent series of Census reports date. By the turn of thecentury the country had rounded out its frontier, had set upits large manufacturing centers. The Census of 1899 in-dicated a working force of 4,713 thousand wage earnersmaking a wide range of manufactured products. The capitalinvested in manufacturing plants was estimated at $8,975million.

Nevertheless manufacturing in 1929 stands in strikingcontrast to manufacturing in 1899. The most evident changeduring the three intervening decades has been in the scale ofoperations. Yet the rise of new products, the development ofnew techniques, the introduction of new methods have alsogreatly altered the structure of manufacturing. New in-dustries have come into being. Old activities have dwindled

12.2 STRUCTURE OF MANUFACTURING PRODUCTION

in importance or have dropped out of the industrial picture.Such industries as the manufacture of aircraft, rayon, me-chanical refrigerators, collapsible tubes, artificial leather, andaluminum are not listed in the Census of 1899. Indeed atthat time most of these products were unheard of. In 1929,on the other hand, the statistics on hammocks, horseshoes,oakum, and wood carpet were no longer deemed of sufficientimportance to warrant reporting in special industrial groups.Of greater importance for the manufacturing structure atlarge has been the growth of certain industries, the continueddecline of others.

We need hardly go back to the time of Hamilton and thecountry's first report on manufactures for evidence of thechanging character of manufacturing operations, eventhough the comparison is most striking when made with thatearly record. Permanent shifts in emphasis as well as tenta-tive adjustments are continually taking place within the ag-gregate of manufacturing activities. Yet these changes onlymodify, in varying degree, what we believe to be more or lesspersistent attributes of manufacturing operations.2 For inlarge part the characteristics of a dynamic economy arestructural, not haphazard, transient relationships. They arecharacteristics of a manufacturing system that has been builtup over a long period; they are slow to change because of thestabilizing influence of human institutions, the stability of aslowly altering fund of capital and knowledge, and thepeculiar and unyielding features of particular manufacturingoperations.2 At various points in Ch. 11 and III some indication of changes in the differentrelations studied has been given. Dr. Kuznets' estimates of the value of the end-products of manufacturing (cited in Table i) suggest that measures relating to thedisposition of productive resources are subject to considerable alteration duringperiods of severe recession. The 1919 and 1929 ratios are almost identical, however.Indeed, the stability of the percentage apportionment between capital formation andconsumers' outlay (at a later stage than that to which the capital-consumptiongoods ratio we have examined relates) during this full period is commented uponby Dr. Kuznets in the presentation of his measures (National Income and CapitalFormation, 1919—1935, National Bureau of Economic Research, 1937, pp. 46—7).

The few comparisons made of the relative importance of productive factors for1929 and other years (pp. 6o, 73, ioo) indicate an equally high uniformity.

SUMMARY

The manufacturing structure might be described in variousways, with varying emphasis upon its different aspects.3 It isnot our purpose to explore the concept. We simply try tosuggest certain characteristics of manufacturing activities bymeans of a cross-section picture. It is as if from a reel ofmotion picture film a single still photograph had beenabstracted. For we have taken, at what amounts to an instantin time, a 'still' photograph of a dynamic economy. We havedone this in order to determine, for at least one short period,the various interrelations and characteristics of a most im-portant type of productive activity. We presume no staticequilibrium on the basis of our findings. We simply say thatat the brink of the greatest recession in business activity ofwhich we have record, this is what those engaged in manu-facturing were doing. We present, as of 1929, no diagnosisof incipient ills. We believe that the persistent relationshipsof the structure we would describe have a life span longerthan the average business cycle. To the degree that ourmeasures reflect these relations they may contribute to thebetter understanding of all business cycles, and not alone ofthe major collapse that followed so closely the period towhich they relate.

To two closely related aspects of the structure of manu-facturing production attention has mainly been devoted: (i)the allocation of economic resources to the manufacture ofdifferent classes of goods; (2) the relative use of differentproductive factors. A complete cross-section view of themanufacturing structure might touch also on the extent ofcorporate ownership or the regional distribution of manu-facturing activity. Problems of personnel, of labor relations,and of character of work might also be considered. However,objective criteria for the analysis of these and other at-tributes of manufacturing are not always easily determined.

See for example the diverse meanings given to 'structure of production' by F. A.Hayek (Prices and Production, London, George Routledge & Sons, 1931, p. 35)and E. A. G. Robinson (The Structure of Competitive Industry, Harcourt, Brace,1932).

12.4 STRUCTURE OF MANUFACTURING PRODUCTION

Moreover, certain of these phases of the problem have re-ceived detailed examination elsewhere. It has seemed well tocenter our attention on the two maj or problems stated and toutilize only readily available data that give promise of im-proving our knowledge of the character of manufacturingproduction.4

Rather than attempt to summarize in this concludingchapter all the detail outlined in the preceding chapters andthe nine appendices, we shall state certain general proposi-tions based on these materials, and endeavor to indicate theirsignificance. Obviously such a method of review leaves muchto be desired; it leaves unmentioned a great deal that thecareful reader will have found by close scrutiny of thevarious tables. The following paragraphs, however, point tomajor findings.

A high percentage of the manufactured goods produced in1929 was destined for human consumption. Of the aggregatesales of manufacturing establishments in 192.9, somewhatover 70 per cent is identified with consumers' goods. In termsof value added, the ratio is exactly 70 per cent. For the morenarrowly defined group of consumption goods, exclusive ofany part of producers' supplies or construction materials, thepercentages are 6o and respectively. Unfortunately wehave no ready means of measuring 'output' other than inthese value terms, and the 'contribution' of the manufactur-ing enterprise is perhaps not well appraised by such marketvalues. Moreover, we are not sure of the extent to which theflow of capital goods is diverted to replacement of existingcapital or spent for expansion purposes. Nonetheless, greatsignificance attaches to this ratio between consumption andcapital goods, particularly in view of the greater cyclical

The data with which we have worked are so detailed and so voluminous that somecondensation has been necessary. As a result, our analysis has only infrequently beenconcerned with the description of particular manufacturing processes. If the realismgiven by titles of individual industries and processes is missing, it is because the scopeof the investigation and the desire to isolate general relationships have subordinatedthe individual manufacturing activity. In Ap. I and II various measures relating to the326 industry divisions recognized by the Bureau of the Census are presented.

SUMMARY 125

amplitude of the fluctuations in capital goods production.But it is just because of these different responses to changesin business activity that the 1929 proportions between thevalues of manufactured capital and consumption goods can-not be taken to define a persistent relationship. The generalmagnitudes of the different operations are evident, however.Since i 929' was on the whole a year of expanded businessactivity, the figure of 70 per cent of the total that we as-sociate with consumption goods and related products is in allprobability a lower limit of the measure than would befound in a less prosperous year.

The four major divisions of consumption goods industries,in order of magnitude of value added by manufacture, arewearing apparel and personal equipment, foods, householdgoods, and motor vehicles. Together the value added inthese four groups of industries totaled 8o per cent of thevalue added for all consumption goods in 1929, and 45per cent of the value added by all manufactures.

Goods with a relatively long service life (over two years)made up 43 per cent of the total value of products, 47 percent of the total value added by manufacture. Roughly two-thirds of these durable goods were capital or constructiongoods; the other third being destined for consumption use.Transient goods, including all producers' supplies, con—tributed approximately one-third of the total (36 per cent ofgross value of products, 33 per cent of value added). Thethird group, semidurable goods, made up per cent of bothgross value of product and of value added.

Most manufacturing was done in plants whose productswe have classed as 'finished', that is, the products were soldfor purposes other than further manufacture. The extentof intermediary fabrication (i.e., production of unfinishedgoods relative to the production of finished goods) wasgreatest in the capital goods group and in the consumers'goods transportation subgroup; it was least for certain typesof consumption goods, particularly manufactured foodstuffs.

126 STRUCTURE OF MANUFACTURING PRODUCTION

As a group it would appear that durable goods are to arelatively large degree manufactured in plants at least onestage of manufacture removed from the ultimate users.While it is true that integration of ownership bridges somepart of these necessary transfers, it remains significant that inthe production of durable goods there are relatively moreopportunities for industrial maladjustments. The point iseven more suggestive when we observe that of all groups ofmanufacturing industries those that make durable goods(particularly durable capital goods) call relatively morelabor effort. These durable goods industries employ roughlyhalf the wage earners in manufacturing, which is con-siderably above the corresponding proportion for value ofproduct (43 per cent). When considered with reference tothe wide cyclical amplitude in the output of durable goods,these factors serve to emphasize the strategic importance ofthe durable goods industries in any explanation of businessfluctuations.

For the most part, manufacturing involves the processingof nonagricultural materials. Value added by the fabricationof farm and animal products was but little more than one-fourth of the total value added by manufacture in 1929.On this criterion, the most important materials were metals,whose fabrication accounted for 38 per cent of total valueadded. The iron group alone accounted for 31 per cent. Thefabrication of nonfarm materials required twice as manywage earners as did the processing of farm products. in termsof their value as they enter the first stage of manufacture,however, farm materials are unquestionably more important.Only when we examine the extent of their processing in themanufacturing industries does their importance in the manu-facturing process shrink. Our estimates suggest a sixfold

We have no direct estimate of the value of products of finished or unfinished durablegoods. All capital goods are durable by definition, and most of the consumers' trans-portation goods are durable also. Construction materials are also durable in their finaluse and, while for the most part they are here classified as 'finished' because they re-ceive no further fabrication in the manufacturing system, are destined to undergoconsiderable change in form beyond the factory door.

SUMMARY 127

increase in the value of minerals, a doubling in the value offarm products.

Measures of the apportionment of various factors inmanufacturing are quite similar from group to group, andfactor to factor.° Subject to some variation, approximatelythe same proportion of different productive resources is re-quired for the manufacture of goods of the types we havestudied. This observation holds whether our measures areexpressed in dollars or in physical units. The general simi-.larity of the ratios is to be seen in the comparisons of ChartH.

Of the 8,839 thousand wage earners employed in manu-facturing industries in 1929 we estimate that 4,900 thousandhelped to fabricate consumption goods per cent). Al-lowance for supplementary consumption goods groups (pro-ducers' supplies and residential construction) raises this toabout two-thirds of the total. Relatively to other productivefactors, a low percentage of total wage earners are employedin making consumption goods, a high proportion in makingcapital goods. we conclude that capital goodsrequire a relatively larger labor input than consumptiongoods. Only in the group of personal consumption goods,clothing chiefly, does the relative labor requirement exceedthat for capital goods. Estimates of aggregate man hoursworked in 1929 show no appreciable divergencies from thepatterns dust described.

The relatively heavier 'input' of labor in capital goodsmanufacture holds chiefly for industries making finishedproducts. Among industries whose products are unfinished,those making consumption goods show the relatively greaterlabor investment (relative to value added and other items).It is notable that while durable capital goods as a group callfor a significantly greater proportion of total wage earnersthan of most other items, durable consumption goods do0 In some degree this tendency is enforced by our method of estimate. A commonratio was used to allocate items relating to a particular industry whenever diverseproducts made such allocations necessary.

128 STRUCTURE OF MANUFACTURING PRODUCTIONChart U

DISTRIBUTION ACCORDING TO TYPE OF PRODUCT OF CERTAINITEMS RELATING TO MANUFACTURING OPERATIONS, 1929

— 4•o dO ._• — 0

— o.o 0 0).c E e 2a c C.-v C., " C — S— — C -o -C. '1 0o 0)0 'a • a • 00 U

0) C

: —

C.n — 00 C 40 C .' '

— C.— 'a —'0 0 C C '0 00> LI > tfl U) C I— X

Consumption goods 100

Foods

80Wearing apparel, etc

Household goods 60

Transportation

Other 40Producers' supplies

Construction materials20

Capital goods

0

Unfinished goods 100

Consumption goodsProducers: supplies,Construction materials 80Capital goods

Finished goods 60

Consumption goods40

Producers' supplies• 20

Construction materials

Capital goods0

Durable goods 100

Capital andconstruction goods

Consumption goods60

goods40

Transient goods 20

0

SUMMARY 129

not. Rather, it is in the group of semidurable consumptiongoods, chiefly textiles and textile products, that the propor-tion of total wage earners is large, relative to other items.When we examine wages paid, rather than number of wageearners, however, the relatipn is reversed, for the ratio forthe subgroup durable consumption goods more closely re-sembles the ratio for capital goods. The durable goods in-dustries' share of total wages exceeds the proportion of totalwage earners associated with these industries. In semidurablegoods industries, however, the percentage of total wages isless than the percentage of total wage earners. The averagewage, of course, is lower in the semidurable goods group.

Typically, the number of salaried workers in manufactur-ing industries is roughly i 5 per cent of the number of wageearners. The ratio is somewhat lower for capital goods,particularly unfinished capital goods; conversely, it is higherfor finished consumption goods. There is clear evidence thatthe employment of salaried workers is relatively more fre-quent at the later stages of manufacturing operations. Theproblems of distribution are probably greater at later stages;as products become more specialized with advanced fabrica-tion the need for salaried administrative assistants becomesgreater. There are, of course, certain industries, such as thepublication trades, in which salaried employees comprise alarge percentage of the labor force. Most of these industriesare apparently in the transient goods group, for which theratio we have discussed reaches a peak (one-fourth of totalwage earners).

The composite of overhead costs and profits, exclusive ofsalaries, is in the aggregate approximately one-fourth ofvalue of product. A relatively high percentage is associatedwith consumption goods manufacture, particularly finishedand transient goods; a low percentage is associated withcapital goods, particularly finished capital goods. Thesevariations in elements of costs are defined in later pages; it

130 STRUCTURE OF MANUFACTURING PRODUCTION

is our purpose here only to indicate the divisions of the ag—gregates among the several groups.

The total capital invested in manufactures in 1929 weestimate at $ço billion. Of this, amount, over one-half wasused in consumption goods industries; as much as 70 percent might be traced to the manufacture of products ulti-mately associated with consumption purposes. Fixed capitalfinds its greatest use, relative to the use made of otherproductive factors, in construction materials and producers'supplies; notably less fixed capital was used in the consump-tion goods group. Only in the foods and transportation sup-plies subgroups of consumption goods was there a relativelylarge fixed capital investment.

In the consumption goods group, however, is found thegreater relative investment in circulating capital, particularlyin the wearing apparel subgroup. No division of these capitalitems according to other classification schemes is available ex-cept 011 the basis of the state reports for Massachusetts andPennsylvania. These data suggest that the ratio of capital tosales is slightly higher in the finished goods group. The ratioof capital to number of wage earners is higher for unfinishedgoods.

Heaviest use is made of mechanical power in the pre-liminary stages of manufacture. The amount of aggregatehorsepower capacity of primary movers in manufacturing inindustries making finished goods per cent) is quite dis-proportionate to that of other factors: 69 per cent of value ofproduct, 71 per cent of value added, 67 per cent of numberof wage earners, and even 75 per cent of number of salariedemployees.

The making of construction materials and producers' sup-plies occasions the relatively heaviest power installations. Inunfinished capital goods manufacture (cf. steel rolling mills)the share of the total power capacity identified with un-finished industries exceeds the corresponding share of num—ber of wage earners against 28.5 per cent). On the

SUMMARY 131

other hand, wage earners play a relatively more importantrole in the finished goods group. When we examine thepower requirements of durable goods, we find that onlydurable capital and construction goods have a percentage oftotal horsepower greater than the corresponding percentagesfor other items, such as value added or number of wageearners. Durable consumption goods clearly have no excep-tional need for power facilities. Nor, for that matter, havesemidurable goods or transient products, except as has beensuggested.

On the average, each wage earner in manufacturingestablishments had available, in 1929, power equipment ofroughly five horsepower. But the amount of power availablevaried widely from industry to industry. Per capita horse-power capacity averaged 3.7 for all consumption goods,ranging from 6.4 for the transportation supplies group to 2.1for the wearing apparel and personal equipment group. Inthe capital goods group 5.1 horsepower per worker wasavailable, in industries manufacturing construction materials7.6, and in the producers' supplies group 8.2. Although thecomparability of the basic horsepower statistics is in somedoubt,7 the figures show with reasonable accuracy the vary-ing importance of power in manufacturing production. Be-cause of the different ways and quantities in which power isapplied, horsepower data are not the best measures of capitaluse. Data based upon the assets of manufacturing corpora-tions are better adapted to this purpose.

Fixed capital per manufacturing wage earner averagedroughly $3,000; in Massachusetts approximately $i,ooorepresented machinery and tools. The fixed capital invest-ment was only $1,544 for each worker in the wearing ap-parel and other personal goods subgroup of consumptiongoods. Capital goods as a group had a per capita investmentof $2,534, construction materials industries $3,994, and pro-

See footnote io, Ch. I, for a discussion of the merts of the Census horsepowerfigures.

132 STRUCTURE OF MANUFACTURING PRODUCTION

ducers' supplies $4,952, the highest for any major group.Total capital per wage earner averaged for all manu-factures as contrasted with the $3,049 per capita estimate offixed capital alone. Approximately the same increase holdsfor each group ratio. The consumption goods group ratiosare increased most, however, on the addition of circulatingand miscellaneous capital items.

The input relations of manufacturing industries are evi—dent in the different elements of manufacturing cost. ChartIII summarizes certain measures relating to elements ofvalue of product, value added, and the composite of over-head costs and profits.

Cost of materials (plus fuels) is the largest item of manu-facturing cost. The over-all percentage of total value ofproduct was 54.9 in 1929, although the median ratio for the326 industries was somewhat less (çj per cent). Materialcosts are a smaller fraction of selling price in industries mak-ing finished goods than at prior stages of manufacture, andin industries making durable goods than in those makingnondurable goods.8 Such costs are relatively low for allcapital goods, and particularly for finished capital goods. Incontrast to 54.9 per cent for all manufacturing industries,material costs in industries making finished capital goods com-prised but 42.6 per cent of value of product in 1929.

Total wage payments were I 5.9 per cent of sales, but themedian ratio for the 326 large and small industries washigher, I 9 per cent. Of the weighted ratios the highest, forthe groups discussed, is that for finished capital goods, 22 per

8 The magnitude of these ratios is determined by the relationship between the valuecontributed in the industry and the value created at prior productive stages. Where themanufacturing contribution is small, as for example in food processing, material costsbulk large in the final selling price. Where the fabrication process is elaborate, andstarts with simple raw materials, material costs tend to be low. So far as industrialintegration results in a lengthening of the manufacturing process in a given industry,the effect may well be to diminish the importance of material costs in the sellingprice. However, the industrial unit of the Census is the establishment, not the corpora-tion, and therefore much industrial integration does not affect the ratios here presented.That there are wide variation8 behind the group averages goes without saying. More-over, the finished and unfinished goods groups are not directly comparable, since manyfinished goods industries cover the entire manufacturing process.

(.fl

C.

r.-

=—

a.cr

n-,

0=

"-4

"—

-a.

oor

—0'

0.0

a.0

0—

(a(a

r —

.—

0.0.

.40'

_-o

—C

,9

C.0

'0s

— in

oS

— (

4r

—a.

a?

— _

c—

C—

—'0

— U

'—

onn'

0,,,

0-,

000

00

5 0

00

00

——

0a.

00

00

0a

0.0

00

00

00

a0

00

00

0.0.

0a

0 0

0C

a.

a

C

aC

C

aa

- (0a

—

II

II.'

.)—

II

II

II

—04

,J

II

I

Os

1I

II

II

II

II

ca11

II

II

II

>Iii

II.

II

0.I

II

II

II

II

II I

0.

-a

134 STRUCTURE OF MANUFACTURING PRODUCTIONcent. The lowest percentages are for finished consumptiongoods and transient goods. Again durable capital and con-struction materials have a common pattern distinct from thatof durable consumption goods. Apparently there is closercorrespondence in the elements of cost between durable andsemidurable consumption goods.

For many purposes, wage payments are properly com-pared with costs of services originating within the manu-facturing process itself, i.e., all costs except materials. Ac-cordingly we turn to measures of fabrication costs as per-centages of value added. Of this total, wages account forabout 35 per cent, salaries i i per cent, and overhead lesssalaries but including profits the rest, per cent. Con-siderable variation around these general figures is observed.

In capital goods industries wage payments average 40per cent of value added. Consumption goods industries alonepay 33 per cent of value added in the form of wages, thoughin the food group these payments fall to less than 24 percent. The ratio rises to 40 per cent in the wearing apparelgroup. The discrepancy between capital and consumptiongoods is less pronounced in the group of industries makingunfinished products, more striking in the finished goodsdivision. In all groups, wages are notably a smaller per-centage of value added in the finished goods division; otheritems, particularly miscellaneous overhead plus profits, be-ing larger.

Salary payments appear to be a fairly uniform percentage(roughly i i per cent) of value added for the different majorgroups of manufactured goods in 1929. In the consumption'goods total, we know of course that they are dispropor—tionately high for the publication group; they are somewhatbelow the typical figure in the foods and private transporta-tion groups. In most instances, there are probably no sig-nificant differences in the salary ratios from group to group.

Overhead costs other than salaries but plus profits, as apercentage of value added, reveals variations complementary

SUMMARY 135

to the measures just described for wage payments, the thirditem, salaries, showing no wide differences. These overheadcosts plus profits are relatively high for consumption goodsindustries, particularly for foods and the items of the trans-portation groups. They are relatively low for the capitalgoods industries. The finished goods group has the higherrelative burden of overhead. It has been suggested thatsales effort on the part of the manufacturer is greater withfinished than with unfinished goods, and if so, the overheaditem would reflect such efforts. For information on the com-ponents of this residual item, typically half as great as thevalue added total itself, we have turned to non-Census data.

Treasury Department statistics on the income and ex-penses of manufacturing corporations in 1929 suggest thatof the magnitude roughly corresponding to the Census item'value added less salaries and wages' about one-fifth repre-sented profits (one-tenth of all fabrication costs and one-twentieth of value of product). Depreciation charges ac-counted for roughly 10 per cent and taxes 6.4 per cent ofthis overhead plus profits item. The bulk of these residualcosts, however, are masked in a miscellaneous item that in1929 was per cent of the overhead item. The Dun-Bradstreet 1933 survey of 1,709 manufacturing concernsprovides measures of certain items of overhead cost thatprobably, though not surely, are included in the Treasurymiscellaneous item. Selling expenses stand high among thedifferent expense items other than cost of goods reported inthe sample survey. This finding supports our belief thatsome costs of distribution are included in the Census over-head item. No great differences in the character of thesecosts are indicated for capital and consumption goods in-dustries.

The significance of these conclusions depends upon severalconditions. In general, the measures of manufacturing opera-tions in 1929 are important because they help to define cer-

136 STRUCTURE OF MANUFACTURING PRODUCTIONtam characteristics of the manufacturing structure. But howwell they do this, how important the areas they define, andhow novel the information they give, are questions yet to beconsidered.

Certain basic measures relating to manufacturing opera-tions hitherto not available have been especially preparedfor analysis in this study. Thus the estimates of averagehours worked in 1929 in different manufacturing industriesprovide new measures bearing on an important aspect of pre-recession factory operations.° The estimates of industry-to-industry relationships in terms of product outlets (Ap. I)have made possible a new set of measures of the net value ofmanufactured goods by groups of industries and of the ex-tent to which this value originates within the manufacturingsystem (Ap. IV). One aim of the survey has been to de-termine these magnitudes—net value, man hours of laboreffort, capital investment—associated with manufacturingoperations in 1929.

Most of the data presented in this volume refer to dif—ferent groups of manufacturing industries. Accordingly, theimportance of the measures rests upon the significance of theclassification to which they relate. The significance of certainof our groupings has been demonstrated by other investi-gators. In general the groups have been selected after con-sideration of their meaning with reference to the source anduse made of productive resources. The data we have collatedrelate to broad industry divisions and are descriptive of thecomplete groups rather than of the individual industries thatcomprise them. The measures refer to selected portions ofthe economic population, as groups.

The significance of the group measures probably variesaccording as they are: (i) measures relating to the relativemagnitudes of the several groups (presented, for the mostpart, in Ch. II), or (2) measures of the interrelations of pro-ductive factors within the groups (cf. Ch. III). The relative

These data are discussed in Ch. I and II, and described at length in Ap. Ill.

SUMMARY 137

group magnitudes are probably less persistent in the face ofchanging business conditions than are the interrelations ofthe productive factors.'° We should expect, for example, thatthe proportions defining the distribution of economic re-sources between consumption and capital goods would besomewhat different in 1933 than in 1929. However, wemight expect fewer discrepancies in the relative apportion-ment of value of product among different items of manu-facturing cost. But to the truth of this hypothesis, only similarsurveys for other intervals can fully testify.

It is in the broader implications of our study, of course,that its true significance lies. While not providing a completedescription of manufacturing operations, or an unequivocalmeasure of the elements of the manufacturing structure, thesurvey bears on a most important question. For it relates to

the most obvious economic problem which confrontsthe inhabitants of any country or of the world as a whole,• . . to determine how the limited natural resources of thecommunity, its limited flow of savings, its limited equipmentof human brains and hands, is to be allocated between theinfinity of different uses in which they are capable of yieldinga harvest of enjoyment"." When D. H. Robertson wrotethese lines he was chiefly concerned with the guiding forcesthat control the disposition of economic resources. This "mostobvious economic problem" we have not considered, but wehave sought to determine in what actual proportions theseresources were allocated in an important area of productiveactivity during 1929.

Some continuing knowledge of how productive resourcesare being utilized is important even in an economy whereadministrative decision does not determine their disposition.It is important for the wise decision of public policy. Theeconomist must consider this aspect of the productive struc-

10 But note the instances of stability in certain of the ratios relating to distribution ofproductive resources cited in footnote z above.

D. H. Robertson, The Control of industry (Harcourt, Brace, 1923), p. 4.

138 STRUCTURE OF MANUFACTURING PRODUCTION

ture in his appraisal of the composite of different productiveactivities, the student of business cycles must keep it in mindin his search for strategic factors in business fluctuations.The economic historian would find in a succession of cross-section studies of the productive system a wealth of informa-tion about the changing structural characteristics of the econ-omy. Since manufacturing comprises an important segmentof the productive system, the present survey of the patternof manufacturing operations contributes to a more completeknowledge of the "obvious problem" of which D. H. Robert-son wrote.

• The peculiarly important role of durable goods industrieshas already been mentioned. Notable is the requirement intheir manufacture of almost one-half of all manufacturingresources in 1929, the more frequent transfers of the un-finished product from industry to industry resulting in in-creased opportunities for maladjustment of supply anddemand, the relatively heavy demand for factory labor andthe even greater relative volume of factory pay roll. Theinstability in the demand for durable goods has increasedsignificance when considered in the light of the relativelygreater influence that their manufacture exerts on the wagereceipts of factory labor and on the entire sequence ofmanufacturing operations.

One of the more striking contrasts revealed in the study isthe relatively small percentage of the sum received for thefabrication of transient consumption goods that is disbursedin the form of wages and salaries. Only 38 per cent of valueadded was paid as wages or salaries in this group in 1929. Inother groups roughly 50 per cent of value added was paidto employees, the highest percentages being in the capitaland construction goods groups. The significance of this re-lationship is obscured somewhat since we cannot trace thedestinations of the payments for items of overhead, andsince a somewhat greater percentage of the original salesprice is paid, relative to other groups, for materials consumed

SUMMARY 139

in manufacture. It would appear, however, that an originalexpenditure on other types of manufactured goods mighthave a more immediate ramification throughout the eco-nomic system. If so, the relationship observed probably oper-ates to retard business recovery and to accelerate the laterphases of business expansion, for the relative expenditureon transient goods (foods being one major component) isgreatest during periods of business depression.

The rather limited significance of the construction mate-rials industries, as distinct from other capital goods in themanufactures total, is not without import. Granted the rela-tively close resemblance to the capital group in the pattern ofcosts and in the use made of different productive factors,the lower absolute level of the totals for construction mate-rials (less than one-half that of capital goods in 1929) setslimits on the significance of the construction industry so faras manufacturing activity is concerned. In size, the con-struction materials group in manufacturing ranks with theautomobile and related products subgroup of consumers'goods.

The growth of the automotive industry and the manu-facturing activities it occasions vividly illustrates the chang-ing structural composition of manufacturing and the chang-ing "harvest of enjoyment". Index numbers have longtestified to the exceptional growth of these industries but itis nonetheless striking to find approximately one-tenth of allwage earners and an equally large percentage of total capitalcontributing, at some stage of the manufacturing process, tothe output of what is relatively a new consumers' service.The operation of the manufacturing system today is coloredby the growth to maturity of the components of this majorgroup of consumption goods industries.

No attempt has been made to establish norms for a satis—factorily operating economy. A rigid scheme of allocation ofproductive resources can hardly be relied upon to provide a

140 STRUCTURE OF MANUFACTURING PRODUCTION

practical basis of control. Studies of the changing pattern ofproduction, however, should further the understanding ofconditions favorable to economic stability. Such studies,covering wide areas of the economic system and relating tosuccessive periods, would seem to be prerequisite to a far-seeing consideration of actual and desirable ways of utilizingthe nation's economic resources. This cross-section view ofmanufacturing production in 1929 is a step in this direction.