distribution of national income · distribution of national income a national income total for a...

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

Volume Title: National Income and Its Composition, 1919-1938, VolumeI

Volume Author/Editor: Simon Kuznets, assisted by Lillian Epstein andElizabeth Jenks

Volume Publisher: NBER

Volume ISBN: 0-87014-039-6

Volume URL: http://www.nber.org/books/kuzn41-1

Publication Date: 1941

Chapter Title: Distribution of National Income

Chapter Author: Simon Kuznets, Lillian Epstein, Elizabeth Jenks

Chapter URL: http://www.nber.org/chapters/c4226

Chapter pages in book: (p. 61 - 95)

CHAPTER 2

Distribution of National Income

A NATIONAL income total for a given country and period canbe apportioned among smaller spatial units or shorter periods.For example, national income for the United States can beapportioned according to its origin in the various states or instill smaller territorial units. An annual total can be appor-tioned among quarters or months to give a more sensitiverecord of temporal changes. But we are not concerned withsuch distributions, largely because our estimates are for thecountry as a whole and annual. Our interest lies instead indistributions that reflect substantive rather than formal char-acteristics. To what uses is the monetary equivalent of national.income put; by what industries is the net national productturned out; what are the attributes of the various factors in theproduction process and the qualities of goods comprising na-tional income?

Three main types of distribution based on these character-istics are attempted in our estimates: (i) among withholdings,disbursements, and consumers' outlay; (2) by industrial origin;

by type of income or payment, representing compensationfor various kinds of productive service. They are discussed herein order to clarify the meaning of the constituents of eachclassification; to indicate the difficulties encountered in defin-ing these classifications precisely; and to explain how exigen-cies force us to depart at some points from the allocations thatwould best serve the purpose underlying the classification.

6i

62 PART ONE

i Among Withholdings, Disbursements, and Consumers'Outlay

In the preceding chapter (Sec. 5 A) we mentioned that betweenthe completion of the production process, whose net yieldconstitutes national income, and the end of the process ofultimate consumption, there are two intermediate phases atwhich we can measure aggregate payments to individuals andconsumers' outlay. Treating these two and ultimate consump-tion as parts of national income, we arrange in the accom-panying tabulation the complementary categories which,together with payments, outlay, or consumption, add up tonational income. The order on the left side is dichotomousfor each of the four larger magnitudes; on the right side it issequential, showing the number of categories (over two) intowhich national income can be divided.

I National Income National Income

i Aggregate payments to individuals i Net savings of corporations andgovernments

2 Net savings of corporations andgovernments 2 Net savings of individuals and en-

trepreneurs

II Aggregate Payments to Individuals Consumers' outlayand Savings of Entrepreneurs

i Consumers' outlay• National Income

2 Net savings of individuals and en-trepreneurs S i Net savings of all enterprises

2 Net savings of individuals

III Aggregate Payments to outlayIndividuals

i Consumers' outlayNational Income

2 Net savings of individualsi Net savings of all enterprises

IV Consumers' Outlay 2 Net savings of individuals

i Ultimate consumption 3 Net changes in consumers' inven-tories

2 Net changes in consumers' inven-tories Ultimate consumption

DISTRIBUTION OF NATIONAL INCOME 63

A INCOME PAYMENTS AND SAVINGS OF ENTERPRISES

Enterprises do not necessarily disburse to individuals amountsequal to the net product originating during the year. A busi-ness corporation may not pay to individuals in wages, salaries,dividends, interest, etc. a sum exactly equal to the difference be-tween the gross value of its product and the value of goods con-sumed in turning out this product. In prosperousyears corpora-tions often disburse less than this difference, retaining somepositive net savings; in poor years they often disburse more, sus-taining negative net savings. Any enterprise whose activity isincluded in national income may have different amounts ofnet income originating and total payments to individuals.

The distinction seems simple, but its application is not. Howshould we treat payments to individuals that cannot be inter-preted as compensation for their services or the services oftheir property utilized in current production? Obviously, suchpayments may be of two types: (i) Enterprises may make pay-ments which, while not in compensation for services to currentproduction, may yet be in payment for past services (or some-times even future). A clear case is that of pensions paid bybusiness firms to their retired employees. (2) Enterprises maymake payments whose connection with past, present, or futureproduction is tenuous; e.g., contributions by business firmsto community chests or other charities, or relief and publicassistance payments by governments.

In either type we can include the disbursements under ag-gregate payments to individuals and estimate the net savingsof the disbursing enterprises after the deduction of these dis-bursements; or exclude them from aggregate payments toindividuals and estimate net savings prior to their deduction.Whichever we do, they are included or excluded under na-tional income as they are or are not paid out of the currentnet value product.

Our practice has been to include such disbursements aspensions, contributions, and relief under aggregate payments

64 PART ONE

to individuals and estimate net savings of enterprises aftertheir deduction. For pensions and similar payments connectedwith past or future services, we can readily justify our practice.It is never too easy to say that any payment is necessarily incompensation for services in current production (consider divi-dends paid by a corporation in a year of greatly reduced ac-tivity, i.e., payment of a discretionary character for the use ofa property that fails to earn a corresponding return); and it ismore practicable to allow for discrepancies in timing betweenpayments and current production. But an even better reasonfor including such disbursements under aggregate paymentsto individuals is that, unlike other net savings (which or-dinarily assume the form of cash, inventory, equipment, orreduction in net indebtedness), they do flow to individualsand are not retained by enterprises; and that we are interestedin all payments by enterprises to individuals, so long as thenexus between the two is one of some sharing by individualsin the production processes of the enterprises.

The case for the inclusion of contributions and relief andpublic assistance payments under aggregate payments to in-dividuals is less clear, although even here some connectionmay be found with past or future services of individuals in theproduction process. Yet the alternative, i.e., to include themunder net savings of enterprises, would yield more misleadingresults. Of course, these disbursements could be omitted fromboth payments to individuals and net savings of enterprisesand treated as an inevitable cost of carrying on the productionprocess. But this treatment would be even more misleadingsince it would understate national income and fail to measureproperly disbursements by enterprises to individuals. For thesereasons, we included relief and similar payments under aggre-gate payments to individuals and estimated net savings ofenterprises after the corresponding deductions.

Our aggregate payments, consequently, include all disburse-ments by enterprises to individuals qua individuals, and sincesome represent compensation for current services and others

DISTaIBUTION OP NATIONAL INCOME 65

have merely an indirect relation to them we must classify allin more detail (Sec. 3).

The distinction between income payments and net savingsseems clear for business corporations, difficult as it may be toestablish statistically. But for governmental agencies and un-incorporated firms it is more complicated. The distinction be-tween payments to individuals and net savings of enterprisesimplies that the gross value of product, minus the cost of goodsconsumed, can be compared with aggregate payments to in-dividuals qua individuals. However, in the case of govern-mental agencies it may be argued that total receipts do notmeasure the gross value of product and that the latter shouldbe measured by some other yardstick. If the cost of govern-mental activity is the yardstick, then net income originatingand aggregate payments by governments to individuals mustbe equal.

In the preceding chapter (Sec. 3 C) we gave our reasons forusing the payment-price basis for valuing governmental serv-ices. If this decision is accepted, the distinction between pay-ments to individuals and net savings by governments parallelsthat.for business enterprises. We mention it here in order themphasize that the application of the dichotomy under discus-sion is dependent upon how governmental services are valuçd.

In an unincorporated firm the entrepreneur is both therecipient of disbursements from it and the man who decideshow much of its net income should be withdrawn and howmuch should remain as its net savings. Can we differentiatebetween net income originating and income payments to theentrepreneur? It may be argued that the latter comprise theentire residue of' net income originating, after payments toother productive factors have been made; that the entrepre-neur whose firm has positive or negative savings is like anyother ultimate recipient of income payments who may decideto reinvest part of his income in the enterprise that employshim or to part of his accumulated savings. If theargument is valid, there is still a significant difference between

66 PART ONE

entrepreneurial withdrawals for consumption (or investmentelsewhere) and the total net income accruing to the entrepre-neur in his firm. But the difference is similar to that betweentotal payments to ultimate recipients and consumers' outlay,and should not be treated as if it were similar to the differencebetween net savings of enterprises and payments to individuals.

Or, it may be argued that savings of unincorporated enter-prises are different from savings of individuals as individuals.For the latter, saving is the result of decisions made with refer-ence to a freely disposable income. An individual receives asalary, wage, dividend or withdraws a certain amount fromhis firm. The payments are usually in the form of freely dis-posable means; and although the income of every individualis subject to many unavoidable drafts, there is a freedomof disposition that is one of the consequences of a fully de-veloped monetary system and a distinguishing feature of anactual payment received. On the other hand, considered as anaddition to the individual entrepreneur's disposable income,the savings, whether positive or negative, of the firm itself arepartly an accounting fiction. The savings that appear at theend of the year may be due to an improvement in accountsreceivable or inventory position; and it would be difficult toclaim that the entrepreneur, after calculating the net incomethat accrues to him during the year, decided to reinvest partof it in additional accounts receivable or inventories, a de-cision similar to one made by an individual investing freelydisposable funds in stocks or bonds. Net savings of unincor-porated firms can, therefore, be viewed as arising from thesame mixture of discretion and helplessness as net savings inmost business enterprises. Consequently, entrepreneurial netincome should be differentiated from entrepreneurial with-drawals, and net savings of unincorporated firms included innet savings of enterprises and excluded from aggregate pay-ments to individuals.

Since it seems to us that the balance of analytical considera-tions is in favor of treating net savings of unincorporated firms

DISTRIBUTION or NATIONAL INCOME 67

as savings of enterprises rather than o1 individuals, we havetried to separate entrepreneurial net savings from withdrawals;although in the present state of the data the measurement ofthe differences is exceedingly rough and tentative. We giveestimates of savings of all enterprises and payments to in-dividuals, excluding entrepreneurial net savings. But since thealternative viewpoint is tenable, we give also estimatessponding to it and allow for it in the classification above. Inthis alternative treatment payments to individuals include netsavings of unincorporated firms; net savings of enterprises ex-clude them; and income payments to entrepreneurs are con-sidered to be equal to the total net income of entrepreneurs,including savings of their firms.

B CONSUMERS' OUTLAY AND INDIVIDUALS' NET SAVINGS

Aggregate payments to individuals represent the means of pay-ment the economic system places at the disposal of ultimateconsumers, constituting their main, but not sole, source ofpurchasing power: consumers may draw upon their accumu-lated assets or use credit to supplement their current income.

Consumers' outlay designates the sum spent by ultimateconsumers during the year on finished commodities and serv-ices. It can be either smaller or larger than aggregate paymentsto individuals. Ultimate consumers, singly or in to to, can spendless than they receive from producing enterprises, realizingpositive net savings and improving their net monetary orclaims position; or they can spend more, sustaining negativesavings and worsening their net monetary or claims position.Since consumers' outlay measures expenditures on finishedconsumer goods, and individuals' savings are the main sourcefrom which capital formation is financed, we divide aggregatepayments to individuals into these two parts.

Two observations should be made about this dichotomy.First, the distinction between consumers' outlay and individu-als' savings rests, in the final count, upon the definition ofultimate consumption and finished goods. If we consider edu-

68 PART ONE

cation as ultimate then expenses for it becomepart of consumers' outlay. But if education is treated purely aspreparation for economic activity and consequently a speciesof investment, expenses for it should properly be treated aspart of individuals' net savings. The conclusion we reached inthe preceding chapter (Sec. 4), that only such consumption isintermediate as represents utilization by business and publicenterprises of commodities and of services of other enterprises,is relevant here. In accordance with it, consumers' outlay com-prises all expenditures by individuals and households on prod-ucts of enterprises except those incurred by the former asmembers of business and public enterprises; individuals' sav-ings are, then, the difference between aggregate payments andconsumers' outlay.1

The second observation relates to the measurement of bothconsumers' outlay and individuals' savings when the imme-diate payment does not cover the full price of the product. Ifa household buys an automobil.e or refrigerator on an install-ment basis, should consumers' outlay for the year include thefull value of the purchase or .only the amount actually paidduring the year? How should purchases on which there was nopayment at all, during the year, but only a corresponding in-crease in consumers' debts, be treated?

The questions are similar to those discussed in establishingthe timing of production (see Ch. 1, Sec. 5 B). But the argu-ment that production is a continuous process and cannot betreated as occurring onl.y at the instant the product appearson the market cannot be applied to purchasing. It would beunrealistic to assert that purchasing is a continuous process,and that a man who buys a car on an installment basis is en-gaged in continuous purchasing during the entire period he ismaking paylTleflts. Nor is it realistic to assert that a purchaseor outlay takes place at the time the payment is made rather

I Since we indudc net imputed rent from owner-occupied houses iii nationalincome, the owner-occupied unit is treated as an enterprise. Purchaseshouses represent, therefore, use of savings, not consumers' outlay.


than at the time the good changes hands and the purchaserassumes th.e obligation to pay. The distinction between con-sumers' outlay and individuals' savings is consequently a moreuseful tool in economic analysis if the outlay includes the valueof all goods that pass from the effective possession of enter-prises to that of ultimate consumers, and if accordingly totalindividuals' savings are scaled down by the value of obligationsthat ultimate consumers may have assumed in purchasinggoods on credit. This interpretation, modified by exigenciesof data, is followed in deriving our estimates.2

C ULTIMATE CONSUMPTION AND CHANGES IN CONSUMERS' IN-

VENTORIES

The services of finished goods purchased by ultimate con-sumers are not absorbed immediately, i.e., they are not im-mediately and exhaustively applied to the satisfaction ofconsumers' wants. The interval between the dates of purchaseand of the exhaustion of services is relatively brief for perish-able goods but substantial for others. If a smaller amount ofgoods is consumed during the year than is purchased, the stockheld by ultimate consumers, viz., consumers' inventories, in-creases. Or, to the extent that consumers' inventories exist atthe beginning of the year, ultimate consumption may exceedconsumers' outlay, causing a decline in consumers' inventories.

Obviously, changes in consumers' inventories must beanalyzed if the structure of ultimate demand and fluctuationsin it are to be understood. Unfortunately, relevant data are fewarid not easy to obtain. Ultimate consumption and changes inconsumers' inventories occur entirely within the householdeconomy; economic study, on the contrary, tends to concen-trate on processes observable in the market place.

In dividing consumers' outlay into ultimate consumptionand changes in inventories, the major task, if we disregard thealmost complete absence of relevant statistical data, is to esti-

2 Such treatment implies that a return to a business enterprise of goods boughtby an ultimate consumer should enter consumers' outlay with a negative sign.

70 PART ONE

mate the current consumption of durable goods. As in the caseof productive enterprises using durable capital equipment, itis difficult to estimate consumption for time units shorter thanthe entire life of the good. Moreover, ultimate consumers, un-like business enterprises, are not forced by necessities ofaccounting and taxation procedure to estimate such consump-tion; and the calculation of current depreciation and obsoles-cence is often neglected even for such goods as a house or apassenger car, let alone other finished goods that represent asmaller outlay.

Another factor that makes it difficult if not impossible toestimate the current consumption of consumers' durable goodsis the luxury quality many of them have, in consequence ofwhich their utilization has a strong flavor of ostentation. Intimes of stress an ultimate consumer may use his car or housemuch longer than otherwise and forego the kudos enjoyed bypossession of a new one. Since durable equipment used bybusiness enterprises seldom possesses such luxury elements, itsconsumption is more strictly controlled by the calculation ofcosts and returns and, as a result, is a more nearly determinablequantity than the consumption of consumers' durable goods.

Nevertheless, consumption and changes in consumers' in-ventories can now be measured for some commodities, and willbecome measurable for more as data accumulate. At presentthe data are too meager for us to estimate consumers' outlayother than as a whole. But there are strong indications that theincreasing prominence of durable goods in the expenditurepattern of ultimate consumers will stimulate the measurementof changes in consumers' inventories.

.2 By Industrial OriginIndustries differ in the raw materials utilized, productionprocesses carried on, and products turned out. Raw materialsdiffer in the degree to which their sources are concentrated ter-ritorially; in reproducibility and susceptibility to technicalcontrol; in exhaustibility and tendencies toward increasing


costs or diminishing returns; in perishability, quality, andmany other technological properties. Production processes dif-fer in the size of the unit that can most advantageously be oper-ated; in the ratio of direct labor and durable capital equipmentto raw materials consumed; in the extent to which physicaltransformation of raw materials takes place; in the temporalcontinuity of operations; in the relation of vital phases of theprocess to skill, etc. Completed products of industries differ intheir distance from the stage at which they are ready for ulti-mate consumption; in perishability over time and in durabilityin the process of use; in the primacy and urgency of the ul-timate needs they satisfy; and so on through the various physi-cal characteristics of products distinguishable according totheir final use.

Superimposed upon these purely technological character-istics of materials, processes, and products are the peculiaritiesof social and economic organization, which also differ funda-mentally from industry to industry. Some industries havemainly country sites; others are perforce concentrated in bigcities. In some industries numerous small unincorporated en-terprises predominate; in others, entrepreneurs do not existand control is concentrated in huge semi-public or public cor-porations. In some industries competition among enterprisesis fairly effective, in others there is no competition. In someindustries overhead costs are minor compared with directcosts; in others, the opposite is true. Some industries cater torecipients of large incomes; others depend upon the mass de-mand of moderate and iow income groups.

This combination of differences in technological character-istics and social and economic organization is an importantdatum in the understanding and measurement of economicphenomena. If measurement is to be helpful in revealing thefactors that make for change and the way economic changestake place, study of how industries differ is indispensable:the technological characteristics of materials, processes, andproducts and the peculiarities of economic organization spell

72 PART ONE

differences in response during both short and long periods.If an appraisal is involved, as in national income measurement,industrial differences must be kept in mind, for they reveal theareas in which the common yardstick used may have some-what different meanings, and suggest the group compositionof the body social for whose satisfaction national income isused.

A NATIONAL INCOME

The distribution of national income among industries is ofthe net value originating, not of the total value of an industry'sproduct. For some purposes, such as estimating waste involvedin an 'unproductive' industry, this is a disadvantage since weneed the gross value. Net income originating in various indus-tries may be interpreted as the contribution of each to thecommon pooi of goods we call national income; or may be con-sidered a measure of the cost to society of the activities carriedon by each. Both interpretations are applicable to incomeoriginating in any singl.e industry, and care must be takennot to switch, without good reason, from the one interpreta-tion for one industry to the other for another industry.

Interest in ascertaining how much various industries con-tribute to a given national net product or how much they claimin compensation for their activity stems largely from the dif-ferences in their activities. An increase or decrease in nationalincome arising from a corresponding change in the net valueoriginating in agriculture is not open to the same interpreta-tion as an equal change in national income attributable to anincrease or decrease in the net value originating in finance.

If an industrial distribution is to provide bases for properunderstanding, each category in it must be well defined. Simi-lar productive activities should not be included under differentindustrial, divisions; no divisions should include essentiallydifferent activities; the classification should be complete, i.e.,not exclude activities of some importance in the economicsystem; and should contain no false categories that would give

DiSTRIBUTION OF NATIONAL INCOME 73to purely transfer and auxiliary functions the semblance of aseparate industry. To establish an industrial classification freefrom such defects is exceedingly difficult because (i) diverseproductive activities are carried on in one operating or busi-ness unit; (2) productive activities and purely ownership func-tions are carried on in one business unit; the productiveactivities carried on by operating and management units osten-sibly belonging to one and the same industry change.

i) One operating unit often carries on diverse productiveactivities, e.g., both extractive and manufacturing activities orboth manufacturing activities and trading functions; or com-bines the production of commodities, transportation, andpower with construction. The activities are significantly dis-parate. Mining is different from manufacturing in that locationand exhaustibility of natural resources are more vital to theformer than to the latter. The functions performed by steamrailroads and other public utility industries are continuous;construction is seasonal. ManuFacturing is concerned withchanging the form of commodities; merchandising, with theirdistribution. Yet within operating units, i.e., within the estab-lishments directly engaged in production, a mixture of theseproductive activities is common. To allocate the net incomeoriginating in the enterprise among the different productivefunctions is a task obviously beyond the powers of a nationalincome investigator, since it would require exceedingly de-tailed cost accounting and some arbitrary allocation of jointcosts.

Business or management units carry on diverse productivefunctions even more commonly than operating units. An en-terprise, whether incorporated or unincorporated, may corn-prise several plants, offices, agencies, which are often engagedin production of different types, though usually complemen-tary to one another. Certain components of net income origi-nating in the enterprise—property income, overhead salariesand, of course, net savings—are ordinarily not allocated to thevarious operating units, but are given for the enterprise as a

74 PART ONE

whole. Consequently, in distributing them, and hence nationalincome, by industries, we encounter a mixture of productivefunctions within larger business units even more frequentlythan within operating units.

2) Enterprises can act not only as producing entities but alsoas ownership units. A corporation may receive dividends andinterest from other business enterprises, and in turn pay divi-dends and interest to them. Do the payments received repre-sent compensation for goods produced by it or payments toit as an investor, i.e., as a possessor of property and funds itdoes not itself utilize? If the payment is to the enterprise asa producer, the net income to which it gives rise may be con-sidered to originate in the receiving enterprise. But if thepayment is to the enterprise purely as an owner, the net incometo which it gives rise obyiously originates in the paying en-terprise.

This distinction demands that we answer two questions whenwe attempt an industrial distribution of a national incometotal. First, how shall we treat enterprises obviously engaged inproductive activity but still deriving part of their gross revenuefrom ownership, i.e., receiving income on investments? In thiscase income in the form of dividends and interest originates inthe paying enterprise, not in the receiving. Accordingly, inestablishing net income originating in a given economic unitwe must subtract from its gross receipts not only the cost ofgoods consumed but also the part of the gross receipts thatrepresents compensation for pure ownership. So far as data areavailable on inter-enterprise receipts of dividends and interest,we can and do follow this procedure.

Second, how shall we treat enterprises that are largely owner-ship units and in which the share of productive activity intotal income may be relatively small? Savings banks and in-surance companies, which are engaged primarily in placingthe accumulated savings of individuals at the disposal of in-dustry, are good illustrations. In the process of mobilizingindividuals' savings and selecting the place for investment,

DISTRIBUTION OF NATIONAL INCOME 75they produce net income, i.e., the net value of the services ofindividuals and capital engaged in them. In addition, theyearn net income, part of which may be retained temporarilybut the bulk of which goes to policyholders and depositors.Can interest paid to depositors by savings banks and the netsavings of the latter be considered as arising in the bankingindustry? Or the payments to policyholders by insurance com-panies and the net savings of the latter as arising in the insur-ance industry? A similar question can be raised about paymentsto investors by all institutions engaged in placing idle balancesat the disposal of productive enterprises.

Unless the industrial distribution of national income is tolose most of its meaning, it cannot be so applied as to attributenet savings and payments to depositors by banks, insurancecompanies, etc. to the banking and insurance industries. Ifthis is done, what is to bar an interpretation of investment asan industry, and of dividends and interest received by wealthyfamilies (who may have formed a personal corporation) as in-come originating in the 'investment' industry? We must recog-nize the possibility that payments may be transferred from onegroup of enterprises to another, and that a given group ofenterprises may be, with respect to some of the income streamspassing through them, not much more than an association ofindividuals in their capacity as investors and ultimate incomerecipients.

Hence, income originating in such industries as savingsbanks and insurance is confined to the net value of the servicesof individuals engaged in them, and excludes payments todepositors and policyholders as well as the net savings of theenterprises. These payments and savings are treated as origi-nating in the industries from whose stocks and bonds theenterprises receive their revenue.3

3) The effect of carrying on diverse productive functions

3 This does not dispose of some technical difficulties introduced by the absenceof relevant data or of questions arising in connection with the treatment of rent.The latter are discussed in Section A; the former, in Part Four.

76 PART ONE

within operating or business units and of combining produc-tive activities with purely ownership functions within oneenterprise would be reduced appreciably if they always re-mained the same. But they are susceptible to change. Changesin technology make for shifts in the relative importance ofvarious types of productive function within one operatingunit. Shorter term cyclical fluctuations lead to fluctuations inthe distribution of the working personnel and of the activetime a single plant devotes to turning out the finished productand auxiliary operations such as repairs and construction; orto production and merchandising. The housing of several op-erating units within one enterprise is the result and the essenceof the process of industrial that developed rapidlyduring the last decades of the nineteenth century and is stillproceeding at a fast pace. Inter-enterprise payments of divi-dends and interest arising from interlocking ownership areanother facet of the same process. And the extent to whichindividuals place their savings with insurance companies, sav-ings banks, and similar institutions is also changing. Sincecomplete adjustment is impossible, the industrial distributionis not precise and the blurred area changes from one periodto anotiTler.

Moreover, comparisons over time must also take account ofchanges that may have occurred in the productive activities

classified as belonging to one and the same industry. Thefunctions of retail trade, professional service, government, andmany other branches of the productive system are quite dif-ferent today from what they were fifty years ago. The nameremains the same, but activities subsumed under it change,without necessarily reducing or increasing the mixture of typesof productive activity.

Any distribution of national income by industrial originis thus subject to serious qualifications. In the nature of thecase, it cannot be accurate for clearly demarcated functionaltypes of productive activity. At best it is a distribution amongthe institutional categories designated as industries, and its

DISTRIBUTION OF NATIONAL INCOME 77interpretation must be qualified accordingly. For example, ifthe estimates show that manufacturing accounts for x andtrade for y per cent of national income, this does not meanthat x measures accurately the share of activity concerned withtransforming commodities, or y the share concerned with theexchange and distribution of commodities among enterprisesand between enterprises and households. Some trading func-tions may be performed by 'manufacturing' and some manu-facturing activity by 'trade'. Similarly, an increase in the shareof trade and a decrease in the share of manufacturing does notnecessarily mean that distributive activity contributes or claimsan increasing share of the national product, and manufactur-ing activity a declining share. It may well be that certaindistributive functions formerly carried on by manufacturingenterprises and included under manufacturing have beenshifted to wholesalers and retailers, thereby swelling net in-come originating in trade. Similarly, an increase in the shareof inconie originating in governmental activity does not neces-sarily mean. that either the price or quantity of governmentalservices proper has increased relatively to the price or quantityof goods provided by the private business system. It may meanmerely that governmental agencies have taken over some ac-tivities formerly pursued by private business, or• have beenforced into new activities.

Since the shifts and overlapping are minor in comparisonwith the persistent and significant differences among catego-ries, these limitations do not render an industrial distributionof national income worthless. The institutions called agricul-ture, mining, manufacturing, steam railroading, etc., whilecontaining enterprises that combine productive activities ofdiverse types, are largely dominated by distinctive productivefunctions. The element of pure trading is minor in manufac-turing, as is the element of pure manufacturing in trade. Whenreally significant changes do occur, the institutional categoriesalso shift, i.e., new industries are recognized and old industriesdropped. The lack of strict correspondence between the insti-

78 PART ONE

tutional categories and the strictly functional segregation oftypes of productive activity is important merely as a qualifica-tion that should prevent erroneous interpretations of differ-ences or changes in the estimates.

B SAVINGS OF ENTERPRISES AND AGGREGATE PAYMENTS TO IN-

DIVIDUALS

Net savings of enterprises suggest the amount of funds madeavailable for investment without recourse to banks and theoutside money market, or the amount of disinvestment sus-stained. The amounts have a high prognostic value, sinceusually enterprises that enjoy large positive net savings demon-strate thereby their favorable market position and are likelyto expand their activities in the future; while enterprises sus-taining large negative savings will naturally be forced to curtailtheir activities. What is true of enterprises is, somewhat lessdirectly, true of industries. Therefore, the industrial distribu-tion of net savings of enterprises reveals one of the prime fac-tors making for changes in the relative importance of variousindustries in the country's total. It is subject to the same qualifi-cations as a distribution of national income by industrialorigin.

Somewhat similar reasoning can be applied to suggest whywe allocate aggregate payments to individuals by industries.So far as changes in total payments differ from industry toindustry, the analysis of aggregate payments must rest uponits distribution among industrial branches. The use of anindustrial distribution of aggregate payments is thus coor-dinate with that of national income, and is a1so subject to allthe limitations discussed above.

A distribution of aggregate payments to individuals by in-dustrial origin may serve also to demarcate groups in the bodysocial. Most of the people who derive their income from agri-culture reside in the country and pursue a mode of life quitedifferent from that followed by people attached to other in-dustries and dependent upon income payments originating in

DISTRIBUTION OF NATIONAL INCOME 79them. Similarly, miners, employees of manufacturing enter-prises, people engaged in trade or in professional pursuits, etc.form fairly distinct social groups. Since the pattern of expendi-tures and savings may well differ for each group, an industrialdistribution of payments would assist not only in understand-ing the background of conflict and cooperation within at leastone set of social groups but also in analyzing changes and dif-ferences in the division of payments between consumers' outlayand net savings and the apportionment of consumers' outlayamong finished goods of various types.

A distribution of income payments by industrial origin is,however, merely a rough approximation to what is wanted andfails to conform to two essential conditions. If our interest liesin the income of various social groups differentiated largelyby their industrial attachment, we should estimate total in-come payments received by each, and exclude income paymentsthat cannot be interpreted as receipts by members of a groupwith a given industrial attachment. For example, by estimatingincome payments originating in agriculture we account forthe major part of farmers' current income and segregate thepart of aggregate income payments that is received primarilyby a social group called farmers. But total income paymentsoriginating in agriculture are both too small and too large forour purpose: too small because they cover only paymentsfarmers receive from agriculture and exclude payments farm-ers receive as compensation either for direct services to otherindustries or for property invested otherwise than in agricul-ture; too large because they include payments not only tofarmers but also to individuals who have little connection withagriculture and do not depend upon it for their income, e.g.,holders of farm mortgages. For any industry the amount ofpayments originating tends to differ in these two respects fromthe total income of the group attached to it.

The industrial allocation of property income, dividends, in-terest, and to some extent rent is especially difficult to interpretin terms of social groups, for the recipients do not actively

8o PART ONE

participate in the industries that are the sources of this income.When merely a minor portion of the income of individualsor households is derived from property, where it originateddoes not help in classifying recipients by social groups. Whenproperty is the major source, receipts are likely to come fromdiverse industrial sources, and dependence upon a single in-dustry is probably uncommon.

The industrial distribution of aggregate payments to in-dividuals is more significant when applied solely to paymentsthat represent compensation for direct services. The industrialdistributions of wages, salaries, and entrepreneurial incomereflect the apportionment of the proceeds of industry to groupsdifferentiated by their industrial attachment. But even thisnarrowing of the scope of payments to be allocated still leavesthe distribution merely an approximation. It is not unusualfor an individual to derive income from more than one in-dustry (either from seasonal or part-time jobs or from diversindustrial attachments, common among the professions). Suchinter-industry combination.s are even more common withinhouseholds or economic families, in which one income earnermay be engaged in one industry and another in a differentone. And the household as consumer rather than the individualas producer is the unit by which income receipts of socialgroups are classified.

Despite these limitations, the second meaning of the indus-trial distribution of aggregate payments to individuals is sig-nificant. Payments originating in an industry are a tolerableapproximation to the total receipts of people attached to it;and can be derived in large part from the body of data uponwhich the industrial distribution of national income rests.

3 Of Income Payments by TypeThe classification of income payments by type is based largelyupon differences in the functions performed by the recipients.Differentiation among these functions is based in turn uponwhether the recipient himself engages in the production proc-

DISTRIBUTION OF NATIONAL INCOME 8i

ess or participates solely through his property; upon the direct-ness of his participation, if active; upon the extent to whichhe shares in the management and disposition of the enter-prise's activities; upon the character of his property claims.Applying these criteria yields the usual classification of in-come payments by type—wages, salaries, other income ofemployees, entrepreneurial withdrawals or net income, divi-dends, interest, rent, royalties, etc.

A PAYMENTS FOR SERVICES OF INDIVIDUALS AND OF PROPERTY

The most fundamental distinction is perhaps between pay-ments for the services of individuals and of property. Theformer are based on direct participation in the productionprocess—the commonest form of economic activity, absorbingthe major part of active economic agents' attention, imposinga pattern on the life they and their families lead, and demand-ing at times considerable sacrifice. Participation through in-vestment, the source of property income receipts, does not re-quire similar activity on the part of individuals and is com-patible with extensive participation in other activities. To besure, property investment is sometimes embodied in the indi-vidual's training and skill, and the return for the services ofindividuals contains a substantial element of return for theservices of property. Yet the difference between income pay-ments in compensation for direct activity by individuals andfor the services of their property holds for a wide range ofcomparisons.

One type of compensation for individuals' services may bedesignated labor income; another, entrepreneurial income.The former represents compensation for services rendered byindividuals who have little voice in the decisions an enter-prise makes and can easily be separated from it. The latter in-cludes compensation for the making of all the responsible deci-sions in the management of the enterprise.

This tripartite division into labor, entrepreneurial, andproperty income payments seems at first well represented by

82 PART ONE

institutional categories of income streams to individuals, i.e.,by wages and salaries, entrepreneurial net income or with-drawals, and dividends, interest, and rents and royalties. Butfurther consideration reveals a lack of correspondence.

Entrepreneurial net income, as measured, is what accrues toentrepreneurs after the payment of all production costs. Entre-preneurial withdrawals are the amounts retained by entrepre-neurs for their own consumption and for investment outsidetheir firms. These income payments or withdrawals, as theyaccrue to or are made by groups of entrepreneurs (farmers,miners, retail traders, small construction contractors, etc.) are,from the viewpoint of the functions they represent, a hybrid ofall three types. A majority of entrepreneurs perform actual,physical productive functions that, under a different form ofbusiness organization, are performed by wage earners or sal-aried employees. All entrepreneurs exercise managerial dis-cretion and make the decisions vital to the enterprise bothinternally and in its relations with other enterprises. An over-whelming proportion of entrepreneurs have a net propertyinvestment in their enterprise. The relative importance ofthese three forms of entrepreneurs' participation varies fromindustry to industry; but dividing entrepreneurial income orwithdrawals into labor, entrepreneurial, and property incomewould be so arbitrary as to serve no useful purpose. What thepreponderant element in it is for the country as a whole ishard to say; but since in the group receiving entrepreneurialincome farmers, retail merchants, small construction contrac-tors, and professional people predominate, the category is byand large that of service rather than of property income; andperhaps preponderantly that of labor income rather than ofentrepreneurial income.

Rent raises a somewhat different question. Net rent paid toindividuals is largely for urban real estate, that flowing fromfarm property and other extractive sources being relativelyminor. Recipients of rent typically take a more active part inmanaging their property than holders of stocks or bonds, who


merely draw dividends or interest. Hence the category is acombination of property and entrepreneurial income. If theelement of entrepreneurial activity were substantial enough,a recipient of rent could be classified as an entrepreneur inreal estate rather than as an individual recipient of propertyincome originating in whatever industry pays the rent. Ifthis were done, rent would not be classified as a separatetype of payment but would become an income stream fromone industrial branch in our classification. Alternatively, rentcould be treated as purely property income and its origintraced to various industries, among which residential realestate is one. There may also be intermediate treatments. Forexample, rent from agriculture can be considered propertyincome originating in agriculture, and all other rent, entre-preneurial income from real estate; or all rent can be classi-fied as property income but assigned to real estate.

There is no decisive reason for our choice of the last-men-tioned method. Had we the proper data, we could perhapssegregate the net rent that represents purely property incomefrom that which is compensation largely for entrepreneurialactivity. Since such data are lacking, it seemed best to treatrent as property income, comparable with dividends and in-terest. As property income, rent should have been appor-tioned among the industries in which it originated (similarlyto dividends and interest); but for lack of Continuous data onrent originating in the various industries (except in agricul-ture and one or two other branches) all rent had to be assignedto real estate.

There is little information on royalties with which we couldsegregate, estimate, or classify them definitely as property orentrepreneurial income. They are probably almost entirelya return on property, acquired either through direct mone-tary outlay or an outlay of labor, although some royaltiesimply more entrepreneurial activity than is manifested byrecipients of dividends and interest. The item is so small that

84 PART ONE

it can be disregarded and since it cannot be estimated sepa-rately on a continuous annual basis, we omit it.

In the total income of employees, how should we treat'other income' and to what extent are elements of entrepre-neurial and property income contained in wages and salaries,particularly the latter? 'Other income' is a miscellaneous cate-gory made up of pensions, compensation for injury,4 and reliefpayments. The second alone has a close connection with cur-rent services rendered in the production process. But all threeare based upon a substantial connection with active partici-pation in tIie production process in the past. If we disregardtiming, it is reasonable to describe all three as payments toindividuals for their services rather than for the services oftheir property. In this respect they are similar to wages andsalaries. Consequently we designate them as 'other income ofemployees' and include them in our estimates under the morecomprehensive total, employee compensation.

There is one substantial element of property income inboth wages and salaries, and at least two minor ones in salaries.The important factors in higher rates of compensation tosome wage earners and salaried employees are their educationand training and the scarcity of the natural capacities neededfor the service they render relative to demand. The part ofwages and salaries derived from higher rates of compensationdue to these factors may be considered property income pay-ments, i.e., returns on the investment made in the past in edu-cation and training or on the value of a natural resourcemonopolized by its possessor. The share of wages, and moreespecially, of salaries, that could be interpreted as property in-come is often substantial, e.g., among professional employees.

Of the other elements of property income in salaries thefirst and more obvious is contained in the compensation paid4 The one industry for which we show this item is steam railroads, Pullman,and express. Compensation paid to persons other than employees is included,since it could not be segregated. The amount in question is probably relativelysmall.


for a sinecure obtained by a property investment, rather thanfor any productive activity or oniy in minor degree for serv-ices rendered. A job is awarded to someone who has made oris making a property investment to the benefit of the groups incontrol of the appointment. Theoretically, such payments areon a par with charity contributions by enterprises on the onehand, with dividends and interest on the other. Data do notadmit of their segregation, but their total is probably notlarge.

The second additional element of property income in sal-aries is in the compensation of corporation executives. Whenthe owners of a corporation are also officials, their salaries areidentical with entrepreneurial withdrawals; and to the extentthat entrepreneurial withdrawals include an element of prop-erty income, so do salary payments to the owners of thesepseudo-corporate units. This item also cannot be segregatedbut is probably relatively minor. It is absolutely much largerin big corporations: the executive personnel, though theoreti-cally subordinate to the stockholders, are actually very influ-ential in making decisions and are, to all intents and pur-poses. the entrepreneurs. The main distinction lies not intheir presumed subordination to the controlling bodies, butin the size of the enterprise. Any enterprise that has attaineda certain size must rely for its entrepreneurial functions not onan individual and mortal owner, but on a more powerful andself-perpetuating group of executives.

In large enterprises the necessity of apportioning entrepre-neurial functions among many people makes for gradationsof power among employees, and the point at which a givenemployee ceases to be an entrepreneur and becomes a subor-dinate is often not apparent. The part of salaries that repre-sents compensation for entrepreneurial activity, therefore,cannot be calculated precisely, but it can be approximatedand we must not forget that salaries, which are often treated asrepresenting labor income, include substantial elements ofentrepreneurial income.

86 PART ONE

Thus the distinction between payments for services and forproperty cannot be clearly drawn on the basis of the institu-tionally prevalent types of payment. There are elements ofproperty income in salaries and in entrepreneurial net incomeor withdrawals; and an element of service income in rent. Yetby and large, wages, all except executive salaries, and mostentrepreneurial net income or withdrawals, are preponder-antl.y service income; dividends, interest, and rent are evenmore preponderantly property income. If compensation forpurely entrepreneurial functions is to be distinguished fromother service income, elements of it will be found in salaries,rent, and, to a minor extent, even in dividends; but there is nosingle, institutionally recognized type of payment in whichit is quantitatively predominant.

B WAGES AND SALARIESAs already mentioned, salaries include a more substantial ele-merit of property and entrepreneurial income than wages. Butthis difference is too elusive and variable to serve as the basisfor the segregation by enterprises of these two types of pay-ment. The salary of a filing clerk or of a typist includes perhapsa smaller element of property and entrepreneurial incomethan the wages of many a skilled worker.

Of the various bases on which the two might be distin-guished—proximity to and directness of participation in theproduction process; manual and non-manual character of theservices rendered; training and education required; methodof payment (piece or time); periodicity of payment (hour, day,week, month, year, etc.); size of compensation—none seemsadequate by itself. Some employees participating in the auxili-ary functions of the enterprise are classified as wage earners(e.g., Construction workers in a factory, repair men, watch-men). Some salaried employees seem to perform primarilymanual functions (e.g., multigraph machine operators, drafts-men). The education and training required of many skilledwage earners is not substantially less, and is sometimes more,


than that required of recipients of salaries. Many wage earnersare paid on a time basis; many salaried employees are virtu-ally on a day and hour payment basis; others (e.g., salesmen)may be on a commission basis.

The answer seems to lie in a combination of these criteria,of which proximity to and directness of participation in theproduction process and the manual character of operationsseem to have most weight. These two factors constitute thebasis for other differences between wages and salaries. Themanual character of operations explains the fewer prerequi-sites of education and training that differentiate most wage-earning jobs from most salaried occupations. Directness ofparticipation in the production process explains a piece ratebasis of wages in many industries and the payment of wagesfor time units much shorter than those used for salaries inmost industries. Finally, direct manual participation in theproduction process usually means that wage earners do notengage in administrative and entrepreneurial functions, ren-ders wages largely prime rather than overhead costs, and,demanding as it does few prerequisites of education and train-ing, is a factor making for the lower levels of most wages ascompared with salaries.

The distinction between wages and salaries is not worthmaking in all industries. There is little meaning in it whenthe production process does not involve much manual labor.For example, banks, insurance companies, educational insti-tutions, professional enterprises, governmental agencies, andeven trade, draw no clear-cut line between wages and salaries.The term 'wages' in such industries is confined to the com-pensation of the few employees who perform manual labor(construction and repair men, charwomen, etc.), and appliesto so small a part of total payments to employees that the dis-tinction is not important. Even industries that employ a largeamount of manual labor in extracting, transforming, and trans-porting commodities make the distinction only if, in addition

88 PART ONE

to employees engaged directly in these processes, there is asubstantial group performing administrative, research, super-visory, or entrepreneurial functions. Where, as in agriculture,such a group is absent or exceedingly small because the size ofthe entrepreneurial unit reduces such functions to a minimumand leads to their performance by the entrepreneurs them-selves, the distinction between wages and salaries again is notworth while. For this reason, our estimates segregate wagesfrom salaries for only such industries as mining, manufactur-ing, construction, and steam railroads.

The distinction between wages and salaries in some indus-tries forces us to segregate these industries, one by one. Inothers the characteristics of the functions compensated by thecombined wage and salary or salary payments vary consider-ably. The very factors that force the separation of wages andsalaries in industries like mining and manufacturing make de-sirable an industrial allocation of total employee compensationamong trade, personal service, government, etc.

It would be illuminating also to have wages classified on thebasis of skill and training, occupation in primary or auxiliaryfunctions, method of payment (piece or time), and amount ofcompensation. Similarly, it would be useful to have salariesdivided into their property and entrepreneurial elements;among various types of administrative, supervisory, etc. ac-tivities; by basis and level of compensation; and by the degreeto which they represent prime or overhead costs. Such classifi-cations, with the possible exception of the segregation of com-pensation of corporation officers from other salaries, are barredby lack of data. This lack of data is not accidental: the alloca-tions suggested demand a close analysis of employee compen-sation within each enterprise. Only when the need for such ananalysis is forced upon an enterprise by its own developmentor by the concern of public agencies for the stability of em-ployee compensation (as under social security legislation) aresome of the allocations suggested made.


C DIVIDENDS AND INTEREST

The distinction between interest and dividends reflects thecharacter of the obligations assumed by the paying enterprise.The obligations giving rise to interest are rigidly fixed withreference to the repayment of principal (as in savings banks),or to th.e continuous payment of interest (as in non-redeem-able bonds), or to both (as in practically all bonds issued by

enterprises). The payment of dividends reflects nosuch obligations. Though many business units pursue a policyof maintaining stable dividends in order to remove speculativeelements from the purchase or holding of their stocks, they arein a position to vary disbursements when conditions are mark-edly above or below ordinary levels, and usually do so. By con-trast, even when there is no definite obligation to pay interestthe existence of an obligation to repay the principal is con-ducive to a conservative investment policy and to a temporalstability of in.terest payments. Since dividends fluctuate andinterest is relatively stable we separate these two types of prop-erty. income.

With this basic difference between interest and dividendstwo others are associated. The first is the presence in dividends,but not in interest, of an appreciable element of entrepreneu-rial income. Short term changes in dividends distributed byenterprises reflect the skill with which they have met changingeconomic conditions and fluctuations in the markets. And ifunder entrepreneurial income we include, among other ele-ments, compensation for their success or failure, dividends ob-viously contain a substantial share of entrepreneurial income.On the other hand, interest on bonds, which carry legal obli-gations, can reflect business conditions only when defaultoccurs, a concomitant of that extreme failure of entrepreneu-rial activity that occurs chiefly during depressions.

A corollary aspect of the difference between interest and divi-dends is that they go largely to people in substantially differentincome groups. Dividends represent a return on the more spec-

90 PART ONE

ulative investment, requiring a discernment and knowledge ofopportunities that possessors of small savings usually do nothave, and bestowing a right to participate in the affairs of theenterprises not desired by a small investor or accessible to onewho holds only a few shares of stock. For these reasons the greatbulk of stocks are held by large investors; and at least in thiscountry, by far the major portion of all dividends disbursed toindividuals is received by people enjoying incomes well abovethe average. Interest paying investments, on the other hand,appeal to small investors; and a large share of interest paid toindividuals is received by people with moderate incomes.

While this generalization is on the whole true, two qualifica-tions must be kept in mind. The first and more important isthat the groups of people receiving interest and those receivingdividends overlap considerably, not only because the same in-dividual may receive both, but also because some individualswith low incomes receive dividends and some with high in-comes receive interest. The second is that interest and, to amuch smaller extent, dividends, frequently flow from enter-prises to individuals via some agency such as a savings bank orinsurance company rather than directly. When such agenciesintervene and we Cannot separate the interest flow to them fromtheir payments to individuals, interest does not necessarilymeasure current interest receipts by individuals, and cannot becompared with dividends. The intervening agencies may re-tain a part of the payments originating in the industries proper,or add to them from accumulated reserves; and there may besome disparities between actual receipts by individuals andthe estimated net interest originating in the economic system.

D DISTRIBUTION OF PAYMENTS BY SIZE

The usefulness of the distribution of family income receiptsby size classes cannot be realized fully unless it is supple-mented by a distribution of the size and other characteristicsof the family, of other economic resources at the disposal offamilies, and of the sacrifices incurred in obtaining income.


Yet its value in the treatment of various problems of economicanalysis and in the interpretation of changes and differences innational income totals is great, even when it alone is available.Its contribution to any interpretation of the welfare equiva-lents of the income flowing to individuals from economicenterprises is patent. Its importance in the social conflict eco-nomic activity engenders between the 'haves' and 'have-nots'is equally clear.

To construct this distribution we need to know how muchincome each family received. It is not sufficient to know the dis-tribution by size of each type of payment separately, since manyindividuals and families receive more than one type. The ideal,of course, would be to have the distribution by size combinedwith that by type, cross-classified by industrial origin. Such acombination of characteristics—size, type, and industrial origin—would shed light not only on differentiae with respect to cur-rent income receipts but, by segregating property income,would suggest the existence or absence of additional resources,and would also indicate differences in the standard of living ofrecipients and the sacrifices implied in the process of earning.However, the unit in this distribution would still be the family,not the industry or enterprise in which payments originate.

It is this circumstance that makes the construction of a dis-tribution of payments by size so difficult. As indicated in thenext chapter, most of the continuous, comprehensive data forthis country are for activities of enterprises and industries; andrecords of income receipts of individuals or families have, untilvery recent years, been confined to the small group that file in-come tax returns. Direct and comprehensive information onthe distribution of income receipts by size is not available evenfor a single year, let alone continuously. We must thereforeconsider substitutes and approximations.

The first and most obvious substitute is the distribution bytype of payment. As noted, one criterion of distinction betweenwages and salaries is the size of average payment; and one dif-ference between dividends and interest is that a preponderant

92 PART ONE

part of the former is received by people with incomes muchlarger than those of people who draw the preponderant part ofthe latter. Entrepreneurial income or withdrawals also are thesource of low average incomes, being in the main receipts ofsmall proprietors. Indeed, a great deal of the public interestin the distribution of payments by type lies in its identificationwith a distribution by size and among social groups. Shifts inthe share of wages are interpreted as changes in the shares of thelow income groups; and similar interpretations are applied tochanges in the distribution between service and property in-come, or between entrepreneurial withdrawals and propertyincome.

We have already indicated how crudely an allocation by typeapproximates a distribution by size. We point out here merelythat departures from a true distribution by size are greaterwhen it is among family units than when it is among individ-uals. Among the former there is more opportunity for incomefrom several sources and greater possibility that substantialincomes from a combination of wages and/or salaries will raisea family to a higher income category than it would be in ifclassified by any one source. Moreover, although incomes fromwages and salaries, or interest and dividends, or labor and prop-erty differ greatly, the hybrid category of entrepreneurial in-come or withdrawals is not so clearly separable from the otherswith respect to the income size class to which it gives rise. Insome industries entrepreneurial income payments or with-drawals are akin to wages; in others, to salaries; in still othersthey tend to yield high bracket incomes.

This suggests one more reason for cross-classifying types ofpayment by industrial origin. Differences in average levels ofwages and salaries prevail from industry to industry; and thelike, as just indicated, is still more true of entrepreneurial in-come or withdrawals. Even dividends and interest may differin this respect among industries, because of differences in thespeculative character of the industries and in the net propertyreturn they yield. Thus the multiplication of cells serves to

DISTRIBUTION OF NATIONAL INCOME 93suggest more accurately differences in size categories withintotal income payments to individuals and families. True, anincrease in subdivisions increases the probability that one in-dividual or family will draw income from more than one cell.But the possibly increased overlapping is surely more than off-set by the greater precision with which differences among pay-ments by size classes would be revealed.

4 SurnmaiyFour types of classification may be used in studying the com-position of the national income total: analytical, evaluative,empirical, and institutional. Analytical classifications are basedupon the analysis of economic reality provided by economictheory and its various applications, analysis that attempts toestablish the various factors making for stability and changeand the interrelations among them in determining economicphenomena. Distributions like that among labor, entrepreneu-rial, and property income types, or between monopolistic andcompetitive industries are good examples.

The evaluative type distinguishes categories within which,for substantial or imaginary reasons, the contents of nationalincome are to be evaluated differently. Extreme examples arisefrom violent prejudices or partisanship. If one happens to be-lieve that blondes form a distinct and exalted group, then in-come payments should be apportioned between fair-haired andother recipients. But usually the grounds of distinction aresomewhat more cogent, resting upon a recognition that types ofactivity or groups of individuals should be segregated for sub-stantial reasons of similarities and differences in pattern oflife and consciousness of kind. A good illustration is the distri-bution of income payments among social groups (farmers,urban manual workers, white collar employees, small businessmen, etc.).

The empirical type of classification is based upon categoriesthat have behaved in significantly different ways in the pastand hence should be segregated for the present and future as

94 PART ONE

a help in diagnosing changes in the national total. It is difficultto find an illustration of a purely empirical classification, but itcould be exemplified by the hypothetical case of an investigatorwho classifies data on net values originating in each enterpriseof the country into groups exclusively by the way they changedin the past.

The institutional classifications are based upon the categoriesin which the statistical data come. They follow the divisionsdetermined by the institutional framework of economicactivity as reflected in current statistics, which, while often suf-fering from sins of omission, rarely err by departing from insti-tutional categories. Any book on national income affordsnumerous illustrations of such institutional classifications.

While these four types of allocation differ sufficiently to war-rant their separation, they have a great deal in common. Eco-nomic analysis is not an exercise of the imagination detachedfrom reality, but must consider the institutional framework ofeconomic activity and deal with the institutions that are re-flected in the statistical data. Nor are the factors analyzed with-out influence upon the groups or types singled out in evalu-ative classifications: they are at least one among several sets offactors that differentiate one social group or one type of activityfrom another. Whatever other factors are involved in suchevaluative classifications, a goodly proportion must be reflectedin the data and hence in the institutional classifications. Thepurely empirical allocations also have considerable kinshipwith the analytical and institutional: they are usually basedupon institutional categories characteristic of the statisticaldata in the past, and if consistently observed, give clues to com-binations of factors susceptible to economic analysis.

The investigator must perforce adhere to institutionalclassifications. Purely empirical allocations are few and unreli-able, since adequately accurate estimates of national income donot cover a period or a number of countries sufficient to yieldempirically established distinctions. And in order to assignquantitative counterparts to analytical or evaluative categories,

DISTRIBUTION OF NATIONAL INCOME 95he should either be able to go behind the published summarystatistics and reclassify the original returns from enterprises orindividuals, or be in a position, by expenditure of time and in-genuity, so to readjust the institutional divisions in the pub-lished data as to get good approximations to analytical andevaluative categories. The former opportunity is usually lack-ing and the latter severely limited.

For this reason our estimates present chiefly the institutionalclassifications of industrial origin and type of payment, with asingle and broad analytical allocation among withholdings,disbursements, and consumers' outlay. For the same reason wediscuss in this chapter primarily these three classifications. Theclassification among withholdings, disbursements, and outlay,being largely analytical and having been made by means of ex-tensive readjustments and recalculations, is discussed in orderto show clearly the lines drawn between the various categoriesand the allocation of doubtful items. In the classifications byindustrial origin and by type of payment there was no need toclarify the distinctions made familiar by everyday discourse.The main purpose was to indicate the extent to which theseinstitutional classifications conform to or differ from the ana-lytical and evaluative ones they approximate but with whichthey are often treated as identical. Thus, a distribution amongindustries is not synonymous with that among types of produc-tive activity or among social groups characterized by their in-dustrial attachment. Similarly, the distribution among wagesand salaries, entrepreneurial net income or withdrawals, anddividends plus interest and rent is not identical with theanalytical distinction between labor, entrepreneurial, andproperty incomes. Even though, without actually carryingthrough the analytical and evaluative classifications, we cannotindicate how far the institutional classifications depart fromthem, we must not forget that they do. Finally, we suggestedthe possible combinations of classifications and indicated theiruse in translating the institutional allocations into the analyti-cal and evaluative categories they approximate.

distribution of national income · distribution of national income a national income total for a...

Documents