AN ESTIMAfiON 0-? THE “WELFARE LOSSES " FROM
MONOPLY EN THE AMERICAN ECONOMY
fhesis §ar il'ha Degree of Ph. D.
MECHIG-AN STATE UNNERSITY
David Roy Kamerschen
1964
1HESIS
0-169
Date
This is to certifg that the
thesis entitled
An Estimation of the "Welfare Losses"
from Monopoly in the Merican Economy
presented by
David R . Kamerschen
has been accepted towards fulfillment
of the requirements for
Ph.D. degree in Economics
July 23, 196u
LIBRARY
Michigan State
University
ABSTRACT
AN ESTIMATION OF THE "WELFARE LOSSES" FROM
HONOPOLY IN THE AMERICAN ECONOMY
by David Roy Kamerschen
For a long time there has been substantial analytical agreement
,—- Ww~~
among economists uponwhat are the unfavorable consequences of mono-
W
poly positions. The monopoliesare said to misallocate resources and
,__~—-—vv.
to redistribute income toward the monopolist. However, untilquite
,fl-
”I
recently, the empirical efforts have been meager. Harberger's attempt .
to get some quantitative notion of the magnitude ofthe misallocstion
of resources andthe consequent loss of "welfare" has been, perhaps,
the best and mostwidely discussed of these recent efforts.
‘19 that study, he found the allocative loss from monopolies
quite small-—less than one-tenth of one per cent of national income.
However, the study was based on but a sample of manufacturing corpor-
ations for the 1924-1928 period. It has also been suggested that be
employed some "heroic" assumptions and questionable statistical proce-
dures. In this thesis, we have continued the work started by
Harberger,‘££‘§l., by tracing a more complete and realistic picture of
the malallocative effects of monopoly. The exact hypothesis that was
tested concerned our belief that our proposed theoretical and statis-
tical modifications of the first approximation model would yield
"welfare loss" estimates of a significantly higher order of magnitude
than had been found in previous studies.
David Roy Kamerschen
Since the malallocative effects stem from the difference between
price and marginal cost, we estimated misallocations by assuming con-
stant costs and investigating profit data. By assuming high profits
are monopoly profits-~subject to a number of qualifications, many of
which can be eliminated by a proper choice of data and periods-~we
estimated the loss by computing the divergence of industry profit rates
- -‘O
a
from the overall average. These estimates were based upon IRS Statistics
of Income data for corporations, partnerships, and sole proprietorships
_‘ for the entire economy, i.e., not for just manufacturing. Since the
1956-1957 to 1960-1961 period was one reasonably close to "long-run
equilibrium" and one in which accounting values were not too distorted,
we used it.
We refined the obviously inadequate raw accounting data in a
profits through adjustments for intangibles, royalties and advertising;
‘J(2) by figuring rates of return on average assets rather than on end-
of-year assets; (3) by computing returns on before-tax and after-tax incomes
and for equity and total capital bases.
The actual "welfare loss" was computed by finding the ratio of
"excess" profits to business receipts and converting this into the
Hotelling formula of 351 r1 2qiic1 , where r1 is the percentage divergence
of actual price from cost, q1 the output, and k1 the demand elasticity--
all of the ith commodity. We computed the losses based upon an
elasticity of unity (Herberger's assumption), of two (Schwartzman's
assumption) and, perhaps, more realistic of all, we estimated the actual
David Roy Kamerschen
elasticities for each industry. We tested most of our findings by
regression and/or correlation analysis as we proceeded.
Although our research uncovered a number of interesting secondary
findings and conclusions, our most significant disclosure was the
acceptance of our hypothesis. The most realistic and complete of our
several estimates put the total "welfare losses" at roughly six per
cent of national income. we may conclude from this that the problem
of monopoly now acquires aggregative significance in addition to its
importance in studying particular industries. In short, we found that
monopoly does affect aggregate "welfare" in a significant way through
its effect on resource allocation.
AN ESTIMATION OF THE "NELEARE LOSSES" FROM
MONOPOLY IN THE AMERICAN ECONOMY
BY
David Roy Kamerschen
A THESIS
Submitted to
Michigan State university
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Economics
1964
ACKNOWLEDGEMENTS
I would like to thank the National Science Foundation for making
funds available for me to proceed with this study. I would like to
express my gratitude to my fellow student Phillip Caruso for the
generous donation of his time in various discussions and criticisms
of this work. I would like to thank Michigan State University for the
many free hours of time on the CDC 3600 computer they allowed me and
for help the computer staff gave me in writing programs. I would also
like to express my thanks to the faculty of the Department of Economics,
Michigan State University, especially Victor E. Smith, Paul E. Smith,
Mordechai Kreinin, and Abba Lerner, for their valuable suggestions in
the Economic Theory-Econometrics Workshops at Michigan State. Harry G.
Brainard and Herbert Kish, who served as members of my guidance com-
mittee, were especially helpful in all phases of this study. Arnold C.
Harberger of the faculty of the Department of Economics, The University
of Chicago, was also most helpful and suggestive in a private correspon-
dence. But I must single out Thomas R. Saving as being the man most
responsible for that which is correct and relevant in this thesis, for
his patient reading of the many drafts which I presented to him, and
his valuable suggestions for both the analysis and style in this work.
Lastly, I would like to thank my wife and family for their constant
encouragement. Whatever errors or faulty reasoning remain in this
paper are, of course, the author's responsibility.
ii
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . . . .
LIST OF FIGURES . . . . . . . . . . .
LIST OF APPENDICES . . . . . . . . . . . . .
INTRODUCTION . . . . . . . . . . . . . . .
Chapter
I. A BRIEF REVIEW OF THE MALALLOCATIVE EFFECTS OF
MON 0P OLY O O O O O O O O O O O O 0
II. THE FIRST APPROXIMATION "WELFARE LOSS" MODEL AND
APPLICATION BY HARBERGER . . . . . . . ,.
III. MODIFICATIONS ON THE FIRST APPROXIMATION MODEL.
IV. ANALYSIS OF THE EMPIRICAL RESULTS . . . . .
v. SUMMARv AND CONCLUSIONS . . . . . g. .
BIBLIOGRAPHY . . . . . . . . . . . . . . .
APPENDICES . . . . . . . . . . . . . . .
iii
Page
ii
iv
vi
22
45
72
114
121
125
10.
11.
12.
LIST OF TABLES
Ratio of Assets to Sales in Manufacturing
Corporations, 1947 . . . . . . . . . . .
Profit Rates for Some Aggregative Industrial
Classifications . . . . . . . . . . .
Frequency Dietribution of Profit Rates for All
Business Establishments . . . . . . . . .
Rank Correlation of Unadjusted and Fully Adjusted
Profit Rates by Methods 1, 3, 4, S, 7, and 8 . .
Results of Generalized Least-Squares Estimation of
CoeffiCients O O O I O O O O O O O O 0
Estimates of the Misallocation of Resources. . . .
Estimates of Aggregative ”Welfare Losses" . . .
"Welfare Loss" Correlations: Product Moment and Rank
Industry-by-Industry "Welfare Losses" for Profit
Method I Including All Intermediate Adjustments
Product Moment Correlation of Harberger's and Our
Estimated "Excess" Profits and "Welfare Losses"
Rank Correlation of Harberger's and Our Estimated
"Excess" Profits and ”Welfare Losses" . .
Product Moment Correlations of Two-Digit Concentra-
tion Ratios, Profit Rates, Lerner's Index of
Monopoly Power and ”Welfare Losses" . . .
iv
Page
53
74
76
78
82
86
89
91
97
102
107
110
LIST OF FIGURES
Figure Page
1. "Dead-Weight Loss" from Monopoly . . . . . . . . 12
2. Harberger's "Welfare Loss" Diagram . . . . . . . 24
3. Profit Rates under Rising Versus Constant Costs . . . 25
4. ”Welfare Losses" under Rising Cost Conditions . . . 26
5. Resource Transfers and "Welfare Losses" as a Function
of Elasticity . . . . . . . . . . . . . 32
6. Graphical Derivation of the Hotelling "Welfare Loss"
FormUI-a O O O O O O O O 0 O O O O O 0 36
Appendix
A.
‘B.
8-1.
3.20
..“\‘ 8-3 0
LIST OF APPENDICES
Measurement and Estimation Proceedures . .
Profit Rate Data for the American Economy,
1956-1957 to 1960-1961 . . . . . .
Frequency Distributions of Average Profit Rates for
Corporations, Partnerships, and Sole Proprietor-
Ships 9 O O O O O O O O O O O O O 0
Average Profit Rates in Corporations--Using Before-
and After-Tax Income and Using Equity and Total
Capital Bases, 1956-1957 to 1960-1961 . . .
Average Profit Rates for All Business Establishments
Profit Methods 1, 2, 3, 5, 7, and 8 . . . . .
Elasticity Estimates for the American Economy,
1956-1957 to 1960-1961 . . . . . . .
Upper Bound Price Elasticity of Demand Estimates
Using the Dorfman-Steiner-Telser Advertising
Intensity Approach. . . . . . . . . .
Price Elasticity of Demand Estimates Using the
Lerner-Robinson Approach. . . . . . . . .
Correlation of Elasticity Estimates . . . . .
Industry-By-Industry "Welfare Losses" for Profit
Method IV Including A11 Intermediate Adjustments;
Unadjusted and Fully Adjusted Losses Using
Methods II, III, V-VIII . . . . . . . . .
Industry-by-Industry "Welfare Losses" for Profit
Method IV Including All Intermediate Adjustments.
Industry-by-Industry ”Welfare Losses" for Profit
Methods II, III, V-VIII, Unadjusted and Fully
Adjusted Results . . . . . . . . . .
Two-Digit Value-Added and Employment Concentration
Ratios Based Upon Percentages Accounted for by 4,
8, 20 and 50 Largest Firms in American
Manufacturing, 1958 . . . . . .
vi
Page
126
134
135
138
145
155
160
168
171
174
175
178
187
Appendix Page
E-l. Two-Digit Value-Added and Employment Concentration
Ratios in American Manufacturing, 1958 . . . . 190
I F. Ranking of Industries by Lerner's Index of MOnopoly
\j Power, Zm = (P-MC)/P . . . . , , , . . . 192
vii
INTRODUCTION
The "welfare" effect of monopoly positions has captured the at-
'tention of economists, at least, since the time of Adam Smith. Over
this time, there has come to be substantial agreement among economists
upon what the unfavorable consequences of these monopoly positions
might be. The monopolies are said to misallocate resources and to re-
distribute income toward the monopolist--all of which may result in a
reduction of aggregate "welfare." Given agreement on principle, it
; seems only natural that the next step would be a quantitative study of
\~
the magnitude of the loss. Surprisingly enough, until quite recently,
’1
the empirical efforts have been meager. Fortunately, quantitative
studies of the monopoly problem in the United States have been made
in recent years by Harberger,1 Schwartzman,2 Kaplan,3 and Weston.4
V— V. W
1Arnold C. Harberger, "Monopoly and Resource Allocation," Pro-
‘geedings of American Economic Review (May, 1954), pp. 77-87; "The
Meaéurement of Waste," Proceedings of American Economic Review (May,
1964).
2David Schwartzman, "The Effects of Monopoly on Price,“ Journal
of Political Economy (August, 1959), pp. 352-362; "The Burden of
Monopoly," Journal of Political Economy (December, 1960), pp. 627-630;
"The Effect of Monopoly: A Correction," Journal of Political Ecogg-
.21 (October, 1961), p. 494; "The Economics of Antitrust Policy,"‘ghg
Agtitrust Bulletin, VI, No. 3 (May-June, 1961), pp. 235-244.
3A. H. D. Kaplan, gig Enterprise in a Competitive System_(Wash-
ington: Brookings Institution, 1954).
4J. Fred Weston, The Role of Mergers in the Growth of Large
Firms (Berkeley: university of California Press, 1953).
1
1
Harberger's attempt to get some quantitative notion of the magnitude
of the misallocation of resources and the consequent loss of "welfare"
was perhaps, the best and most widely discussed of these efforts. In
that study, he found the allocative loss from monopolies quite smallv
(less than a.tenth of one per cent of national income).1 However, the
study was based on but a sample of manufacturing corporations in the
1924-1928 period. It has also been suggested that he employed some
"heroic" assumptions and questionable statistical procedures.2]/IDIEEISII 3
paper, we have continued the work started by Harberger,‘5g_gl. by
tracing~ataors complete and realistic picture of the malallocative ef-M
facts of monopoly. This has been done by modifying the basic theoreti-
cal model and by utilising more recent and improved data. To be sure,
even with our proposed modifications, it should be kept clearly in mind
that this is not the kind of task that one can do with great precision. {
V 7+—
1Schwartzman, while employing a similar technique, and Kaplan,
an entirely different one based on the extent of instability in the
relative fortunes of the leading firms, reach the same conclusions.
2The most important critiques on the above findings are by
George J. Stigler, "The Statistics of Mbnopoly and Marger," Journal
pf Political Economy (February, 1956), pp. 33-40, who examined Kaplan's
and Harberger‘s results, as we shall discuss in detail below; Ruth P.
Mack, "Discussion," Proceedings of American Economic Review (May, 1954),
pp. 88-89, who examined Harberger's results and argued mainly that the
loss must be small since total profits constitute only a small propor-
tion of total income; Halter Adams, "Consumer Needs and Consumer Sover-
eignty in the American Economy," Journal of Business (July, 1962), pp.
264-277, esp. 265-266, who examined both Rarberger's and Schwartzman's work
and argued that they only had "negative" value and they should have
Iooked at the "total" optimality conditions frOm the producer's sfide.
Hicks' "total" conditions say that, if welfare is to be maximize , it
must not be possible to increase welfare by prOducing a new product; or
using a factor not otherwise used. .See Melvin Reder, Studies in the
Ihgorx of Welfare Economies (New York: Columbia University Press,
1947), pp. 37-38.
However, in a great many problems, such as the social control of
industry, a feeling for the general order of magnitude would be helpful. 1/
We have reason to believe that a study of this type can be more
than an "intellectual exercise" since, rightly or wrongly, the findings
of Harberger, 55.21., seem to have had a profound impact on both the
general public and the esonomics profession. This is best attested
to by the recent (November-December, 1963) Chase Manhattan Bank
Survey of college and university economists in which one of the ques-
tions and its tabulated reply was the following:1
Does monopoly on the part of U.S. business now constitute:
A minor problem? . . . . . . . . . . . . . 70%
A major problem? . . . . . . . . . . . . . 23%
No problem at all? . . . . . . . . . . . . 72
This appears to be but another example supporting the famous
Keynes quotation at the close of his controversial classic concerning
the underrating of the power of ideas of economists and political
philosophers. Thus, we think that any proposition that has as many
widespread ramifications as the "welfare" problem of imperfectly competi-
tive markets is deserving of more up-to-date and detailed analytical
and empirical study. The exact hypothesis that will be tested in this
study concerns our belief that our theoretical and statistical modifi-
cations should yield a "welfare loss" estimate of a significantly
larger order of magnitude than has been found previously.
1It should be mentioned that although this was not based on a
strictly scientific sample, the survey of academic economists did
cover a broad, unbiased cross-section of American college and university
economics teachers. It is interesting to note that if the term "U.S.
labor unions" is substituted for "U. 8. Business" the proportions become
50%, 44%, and 6%, respectively. We shall have more to say on this
lateru The above mentioned survey was reported in.§usiness in Brief,
Economic Research Department, The Chase Manhattan Bank, New York 15, N.Y.
The empirical efforts in this study shall rely on the theoretical
proposition that the undesirable impact of monopoly on the allocation 5
of resources may be measured by the divergence of price from marginal
cost in different industries. Unfortunately, marginal cost data are
especially difficult to obtain. However, by assuming constant costs
in the relevant range, for the industry, we can utilize the more
accessible profit data to estimate the losses. In fact, under this ..
assumption Lerner's measure of monopoly power, Zm = §_%_!§. (or l-MC/P)
exactly coincides with the ratio of "excess" profits to total revenue
(sales). In other words, this latter figure now tells us by what
percentage prices in each industry are too "high" or too "low" com-
pared with those that generate an optimal resource allocation. ,f’
Our central argument is that we may pick out the places where ;
resources are misallocated by looking at profit rates. Industries, \/
which have higher than average rates have too few resources and those
with lower than average returns have too many resources. To know
exactly how big a shift it would take to equalize profit rates in all
industries, we have to know something about the elasticities of demand
for the goods in question. For in this model, the "welfare losses"
go up when the elasticity of demand increases.
Of course, our central thesis that high profits are monopoly
profits is subject to a number of qualifications. However, by making
certain adjustments in our data, we are able to estimate "excess"
profit rates that reflect primarily the monopolistic elements. As-
suming the desired resource reallocation from "low" to "high" prof-
it industries is effected, we then measure the net gains to society.
we shall be basing the "welfare loss" estimates upon profit rate J
data for all types of industry (not just manufacturing) and for all
types of business establishments (not just corporations). These rates
shall be computed for both befOre-tax and after-tax income and for )I
4”___,..____
F'—
both total capital and equity bases. Furthermore, estimated industry-
¥._’__,,____,
by-induSETyelasticity estimates shall be employed rather than assuming \2
the same elasticity, of one or two, for all industries as some others
have done. lfi
This brief sketch indicates the general approach we shall be
taking in the pages that follow. MOre specifically our format for the
rest of this study is as follows:
In chapter I, we briefly review the general nature of the "welfare
loss" due to monopolies and the efficacy of our index to measure this
loss. Chapter II describes the first approximation "welfare loss"
model and the results of its application by Herberger. Chapter III
indicates the modifications we shall make to render the model more
useful. Included in these modifications are some pregnant suggestions
of Stigler. To avoid undue misinterpretation, the exadt content of
the assumptions employed in both models are spelled out in detail. In
Chapter IV, we make an actual application of the modified model dis-
.cussed in Chapter III and analyze the results. Our method in this
chapter is one involving successive approximations as we proceed from
a simplified model to one as realistic as the data permit. In Chapter
V, our concluding chapter, besides reviewing what has already been
done, we shall speculate on the direction of some of the factors we
were not able to quantify into our analysis. Finally, in the appendices,
we have discussed our measurement and estimation procedures as well as
including heretofore unpublished data on profit rates, rankings of in-
dustries by Lerner's measure of the degree of monopoly power, elasticity
estimates, concentration ratios and our estimated "welfare losses."
CHAPTER I
A BRIEF REVIEW OF THE MALALLOCATIVE
EFFECTS OF MONOPOLY
Before discussing, in detail, our model for estimating the
"welfare losses" due to monopoly, we first went to review something
of the general nature of these "losses." At least since A. P. Lerner's
interesting paper in the early 1930's, most economists have discussed
the undesirable impact of monopoly on the allocation of resources in
terms of the divergence of product price from the marginal (incremental)
1 Without going into the rigorous proofs employed
costs of production.
in welfare treatises, we shall sketch the line of reasoning needed to
establish the optimality condition of price (P) equals marginal cost
(MC). The reasoning proceeds as follows: We know that society will
fi— 1 w—v—v ——7 1f
1Abba P. Lerner, "MOnopoly and the Measurement of MOnopoly Power,"
‘Regiew of Economic Studies, Vol. I (June, 1934), pp. 157-175. Strictly w
speaking, the P = MC formulation is wrong or, at least, misleading. The
actual requirement for optimality is that P = vmf (value of the marginal
quantity of factor which is the physical increment of the factor multi-
plied by the price’per unit paid for it and received by the owner of
the service--if this increment is exactly one unit of factor, vmf will
equal the price of the factor, pf). The concept of mf, the quantity Of j
factor that must be added to produce one more unit of product, being
symmetrical to the mp, the quantity of product that results from applying“
one more unit of the factor. The usual implied assumption of perfect
competition in buying factors (so that MC 2 vmf) is what makes it only
misleading. Alternatively, the optimality "Rule" can be stated as vmp a
pf instead of p s vmf form. But we are neglecting all these refinements,
as well as, the subtle distinction between proportionality vs. equality
of P and MC. The authority on all these points being A. P. Lerner, The
Economics of Control (New York: Macmillan Company, 1944). For a review
of the other requirements necessary to make P a MC a "good" thing (given
value judgments) see I. M. D. Little, A Critique of Welfare Economics
(London: Oxford university Press, 1950, 2d ed.), p. 45.
7
maximize its social etonomic value from the use of its productive
‘rsspurces only if it is unable, by re-allocatigg its resources, to add
more social_value; however. defined, than it destroys (this result is
automatically brought about in a perfectlygfunctionigg competitive
economy, at least, in the Pareto sense which we shall be concerned
with here) Since the only objective method of measuring the relative
want-satisfying power of a good or service is in terms of the 23222
which consumers are willing to pay for it, we conclude that free-market
equilibrium consumer prices reflect consumers? marginal evaluation of
the goods. If Px = $1 and Py = $2, each consumer adjusts his expenditure
so that a unit of Y is wonth to him twice as much at the margig_as one
unit of X--, i.e., MMy = 2 MUx (MU is the marginal Utility of the given
commodity). In equilibrium, Px a MCx, and Py = MCy; hence, in the
above example MCx = $1, MCy = $2. But marginal money costs are a
reflection of marginal social economic costs. Hence, to produce one unit
of Y, society muqt.give up, at the margin, two units of X (Marginal
costs‘equal the sum of additional outlay by the firm on the extra pro-
duotive services required to increase output by one unit. This is the
sum of additional wages, interest, rents, and "normal" profits required
to be paid per extra unit of output. But the amount of wages, interest,
etc., a'firm must pay for productive services is the amount these
services are worth in other uses, i.e., the vmp--mp times the unit
price of the product.) Hence if Px = $1 = MCx and Py = $2'= MCy,
society's relative evaluation of X and Y is the game as the social costs
of producing x and Y. In this case, there is an optimum allocation of
resources,1 for there is no re-allocation which adds more social value
than it would destroy. Unfortunately, in practice, the selling price
may be higher than MC in many industries and firmso-but to varying
degrees. Where the gap is small, the deviations of actual output from
the "ideal" output are likely to be small. A wide gap would indicate
that output most be increased considerably before the gap would dis-
appear.
In summary, monopoly leads to non-optimal resource allocation
because the money price of any product, which is society's index or
measure of relative worth, benefit, satisfaction of a product at the
margin, is not equated to the marginal costs of production,.which
measures the sacrifice, cost, disutility which was foregone in sacri-
ficed alternative commodities to produce another unit of this commodity.
When Px exceeds MCx, this indicates that society values additional units{
of X more than the alternative products which the appropriate resources
could otherwise produce. Thus, there is an underallocation of resources
to this product from society's point-of-view. When Px is less than the
MC): the reverse holds--an overallocation of resources to X. This must
be qualified to the extent P does not measure all the benefits and MC‘
does not measure‘gll the sacrifices involved. Therefore, when social
revenues, such as chest x-rays and polio shots, and social costs, such
as smoke and pollution, exist, P and MC are no longer accurate indices
1Actually to measure all the deviations from the optimum alloca-
tion of resources, the Lerner measure of the "degree of monopoly,"
(PdMC)/P discussed in detail below, must be supplemented by the formula ,9
for the "degree of monopsony." The relative gap between average '2:
cost (AC) and marginal revenue (MR) is the measure of this latter force.g'
10
of satisfactions and sacrifices.1 In other words, when there are
external effects in production or consumption,2 P = MC does not lead
to an efficient allocation of resources. However, we shall neglect
this qualification in our discussion.
A further refinement which we shall touch but lightly is the so-
called theory of "second best."3 The general theorem revolves around
situations in which the Paretian optimum, which requires the simultan-
eous fulfillment of all the optimum conditions, cannot be met. The'
theorem states in a general equilibrium system with a constraint on
one or more of the Paretian conditions, the other conditions, although
still attainable, are, in general, no longer desirable- In other words,
non-fulfillment of one optimal condition means optimum now requires
departure from all the other Paretian conditions. Similarly, it is not
true that a situation in which more, but not all, of the optimum condi-
tions are fulfilled is necessarily (indeed, even likely) superior to a
1 .
There are, of course, other Schumpeterian-type arguments empha- »/
sizing the "dynamic" over these "static" conditions also. Furthermore,
Reder, op. cit., argues this optimal only as viewed from the consumers'
side--given existing products. We should examine the producers' side
to see if there are any new products that consumers would like newly
produced or old products they would like at new prices. We shall ne-
glect all these arguments because of their nonoperationality. In other
wads, it is virtually impossible to quantify these things. Of course,
A. Smith, Pigou, and Kahn were also important contributors and criticizers
of "welfare" economics.
s.
21.6., "that the utility level of a consumer does --- depend upon
the consumption levels of others, and that the total cost of an entre-
preneur does ... depend upon the output level of others," James M.
-~£bnderson and Richard E. Quandt, Microeconomic Theory (New York: McGraw-
‘ Hill‘Book Company, Inc., 1958), p. 212.
38cc James E. Meade, Trade and Welfare (London: Oxford University
Press, ISSS), esp. Chapter VII, pp. 102-118, and R. G. Lipsey and Kelvin
Lancaster, "The General Theory of Second Best,f Review of Economic
Studies, Vol..24 (1956-1957), pp. 11-32, and the ensuing comments.
11
state where fewer conditions are fulfilled. Furthermore, it is not
true that a situation in which the optimal departures are of the same
direction and magnitude is superior to one in which they vary. This
latter fact means that there is no reason to believe a situation in
which there is the same degree of monopoly in all industries is nec-
essarily superior to one in which the degree varies between industries.
However, if MR$MC in one firm, the "second best" conditions require
that the equality be departed from in all firms. Here, as in the case of
the other refinements, we shall be forced to neglect this "second best"
argument except for a very brief extension in Chapter III.
Returning to the "welfare losses" due to monopoly, we should
mention that misallocated resources may not be the whole of this loss. \/
If there are'bxoess" (supernormal) profits earned, there 291.22 an
undesirable or desirable impact on the distribution of income. The
deleterious effect would result if a larger share of the national income
went to people who are less deserving-~however defined. However, this
criticism is on an entirely different level from the previously mentioned
distortion due to the misallocation of resources. This is because it
is entirely possible that the recipients of the enlarged share of income
might be people more deserving, as defined by our cultural standards--
1
whether they be poorer, nicer, whiter, etc. As Lerner has repeatedly
emphasized, it is better to separate the distribution from the allocation
problem. It is also true that we can have P #:MC and have no "excess" w
profits to redistribute in the first place-~the so-called Chamberlin ‘ ,
"tangency solution." Perhaps a better example that monopoly profits are
not the greatest evil of monopoly is when a monopoly firm with
12
horizontal AC and MC curves sets its price above MC. If the state
charges a franchise tax equal to the profit, the misallocation of re-
sources would persist as the lump sum tax does not affect quantity
and hence MR or MC. The government would reap the gain instead of
the firm; but, the consumers would still get underproduction and 2323:
pricing on this product. We shall neglect any possible redistributiona
effects of monopoly and concentrate on the more tangible and more impor-
tant area of resource allocation.1
Perhaps the exact nature of monopoly distortion can be better
visualized with a diagram. In the figure, we shall assume constant
(horizontal) AC and MC as we do in our model. In competitive equili-
brium G, society would be receiving the Marshallian consumer's surplus
equal to the area of the triangle CGA. As the monopolist raises his
price to E in order to maximize profits (i.e., where MC = MR), the con-
sumer loses to the monopolist that part of his consumer's surplus repre-
sented by the profit rectangle qfiss and is left with but BEA. The little
triangle MGE represents the "dead-weight loss" that goes to no one
Price,
Cost '
Bo "Dead-Weight Loss"
‘ . G
c “ AC and MC
A AR
0 ' * ’
I Quantity
J \ MR,
Figure l.--"Dead-Weight Loss" from Monopoly
1This means that when we speak of "welfare losses" we are using
the word "welfare" loosely to denote economic efficiency. V/
13
for the consumer loses more than the monopolist gains in profit. Even
if the profit were recaptured by lump-sum taxation, inefficiency would
still claim the "dead-weight triangle," HGE.
The nature of this monopoly loss being indicated, we might now ‘J/
turn to the problem of trying to measure the degree of monopoly. The
chief difficulty of doing this lies in the fact that monopoly is only
perceptible by its causes or effects.‘ For as is power, strength,
capacity, potential, force, etc., it is not directly measurable. There-
fore, it is not surprising that there have been a number of proposed
indices suggested to measure the degree of monopoly. They include: ‘
,
(l) the relative gap between MC and P, i.e., Zm =-Z—%—!§, for a profit \/
maximizing firm in equilibrium this reduces to the reciprocal of the
price elasticity of demand (Lerner); (2) an adjusted rate of profit
(Bain); (3) indices of concentration and numbers (Monopoly Subcommittee);
(4) the amount of price inflexibility--including frequency and amplitude, f
of change (Means, Neal, Dunlop); (5) the ratio of total gross profit to
total gross receipts or the ratio of gross profit margin to price
(Kalecki); (6) the ratio of the slope of firm demand, "species" demand
curve, to the slope of industry demand, "genus" demand curve, (Roths-
child); (7) the cross elasticity of demand, which is the ratio between (
relative changes in the quantity demanded of the product considered and C
the relative changes in the price of another firm's product (Triffin); i
(8) the coefficients of penetration and insulation, or the firm's
capacity to penetrate rival markets and its capacity to withstand ‘/
attacks on their own share (Papandreou).
14
Since these are all critically reviewed, in detail, by Professor \i/
M’achlup,1 we shall restrict ourselves to a brief discussion of the ones
utilized here'gig,, (l) and (2). The Lerner formulation of the monopoly
power in £2335 (actual) not potential ("intentional"), (P-MC)/P or
1 - MC/P, is probably the most sensible (Zm = 0 in perfect competition
since MC =P, the other pole being Zm = unity [1] if P is infinite
or MC = 0 ,thus, 2m shows the relative deviation from pure competition).
To be sure, it is far from perfect. It has been criticized for failing
to take accountof the "degree of monopsony," for ignoring the non-price
dimension of imperfect competition, and for not taking account of the
existence of substitutes (this objection presumably led to the Triffin /
formula). It has been further criticized for applying only to a single
firm and only with many qualifications to the whole economy (if all
suppliers had the same degree of monopoly power, Lerner feels there
would be no deviation from the optimum--subject to two qualifications
concerning the "degree of monopsony" and the "production of leisure"),
for lack of data and difficulties oftmaasurement, and for failing to take
into account output restrictions not due to imperfect elasticity of
demand.2 This latter point means that there might be pure competition
1Fritz Machlup, The Political Economy of Monopoly (Baltimore: The V/
John Hopkins Press, 1952), Chapter xii, pp. 469-528. For a brief review
of the main ones see H. H. Liebhafsky, The Nature of Price Theory (Home- I’
wood, Illinois: The Dorsey szas, Inc., 1963), pp. 292:297, or G. Malanos,
Intermediate Economic Theory (Chicago: J. B. Lippincott Co., 1962),
Chapter xx, pp. 501-533, especially pp. 514-533.
2Joel Dean, Managerial Economics (Englewood Cliffs, N. J.: Prentice-
Hall, Inc., 1951), p. 108, n. 94, also contends that it is not a complete .L
measure of the social cost of monopoly ". . . since it did not include
the size of the losses of output and distortions of income that would
result, nor did it allow for the pure profits that appear with a new and
growing product."
15
in a certain market while entry into the industry is restricted by
some artificial barrier. Thus there might be no gap between MC and P
but a margin between AC and P, i.e., "excess" profits. Although this
is not too important since imperfect elasticity is often connected
with imperfect entry, the measure of the relative P, MC gap should
be supplemented with knowledge of adjusted profit rates. In essence,
as we shall see below, we have done this. All in all, we feel that
the main difficulty of the price, incremental cost comparison is one
of impracticality. The notion of MC being an especially difficult
concept to quantify. This is attested to by the protracted contro-
versy in the literature, especially since the 1930's, over its al-
leged shape. Furthermore, one might find two different degrees of
monopoly for the same firm depending upon whether one refers to
long-run or short-run considerations. Lerner regards the short-
period as appropriate for his formula.
To avoid most of the difficulties, we may assume constant
costs (horizontal AC and MC). In such cases, the ratio of monopoly
profits to total revenue (sales) coincides with the Lerner index,
since pq = total revenue and we have assumed constant costs
(AC = MC), profit per unit, P - AC, may be expressed as P - MC--
therefore S$:%Q%a reduces to the Lerner formula (taking out the
common factor q and substituting MC for AC), coincides with the
Lerner index. Data of this first kind, dealing with profits and sales,
are much easier to secure. However, it is only under our very special
assumptions that the monopoly revenue to total receipts ratio exactly
coincides with the theoretically more acceptable Lerner formula. In
cases, similar to the "tangency solution," etc., mentioned above, where
16
there is a divergence between the approaches--in this case there are
no "excess" profits but P;t MC--it is the Lerner approach which is
valid. By using the ratio of economic or "excess" profits to sales
revenue, we may find out how much prices are too "high" or "low"
relative to the competitive positions, if we first find which in-
dustries are earning higher than average rates of return on capi-
tal. For example, if the ”excess" profit rate of sales was equal
to 25 per cent of sales, this means that average costs are 75 per
cent of the average price at which sales are being made. Hence,
this ratio effectively measures the ratio of average price to
average cost, which is assumed equal to MC here.1
By adopting this approach, we do not mean to imply an outright
condemnation of all "excess" profits. Short-term "excess" profits,
as are losses, are justifiable and therapeutic if the economy has
booms which alternate with slumps (at least enough to counterbalance
depression losses); as an incentive for an industry to enlarge the
output of a product in short supply; for the firm which is superior
or exceptionally efficient vis-a-vis its rivals.
1For larger accounting profits on equity to indicate larger
excesses of price over AC, it is also necessary to assume roughly
equal capital turnover.
2For a detailed discussion of this see Joe S. Bain, Industrial
Organization (New York: John Wiley and Sons, Inc., 1959), pp. 371
ff. Of course, persistent, prolonged, or chronic ”excess" profits, h
over a long period of years, must be judged somewhat differently.
"Excess" profits refer to any return greater than “normal." "Normal”L
profits being defined as equal to what the entreprenuer could obtain
with his capital if he used it in some other way, less an allowance
for the inconveniences of transferring it, and plus (or minus, if‘
risk preferences prevails in the community over risk aversion) allowu
ances for any non-monetary advantages. Of course, the shorter the
period of time considered, and the less mobile the capital in pro-
cess, the less will be the "normal" profits (and "total costs").
17
It is for this reason that in our empirical search we shall
attempt to find periods which roughly represent "long-period equili- V\
brium." But even doing this may not be enough. Some would argue 1‘
that divergent profit rates also occur from the dynamics of growth and
development and would be forthcoming even under competitive conditions. \/
A complete model would take all these things into consideration.1
The fact is that we can think of a number of causes of super-
normal profits besides monopolistic or monopsonistic restriction of.
output: (1) windfalls from misestimation of future demand or cost or
lagging adjustment to changing demand or cost--in more general terms,
Knightian uncertainty; (2) the fact that reported profit statistics
often contain elements of return which are really igplicit factor
returns due to the natural scaracity of specific resources,e.g., the
accounting profit which is really rent from the superior ability of
expert management; (3) the riskiness of enterprise investment in
various lines resulting in the payment of "risk rewards" to successful.
risk-takers, losses to unsuccessful gamblers; (4) the rewards of
Schumpeterian type innovation or enterprise.
Despite all these qualifications, we can still roughly identify
monopoly power with high rates of profit.3 Although it is empirically' j,
1Harberger used 1924-1928 as an approximation to'iong-period
equilibrium'and arbitrarily allocated one-third of profits to monopoly J)
profits.
2Although some schools of thought would put all of these sources
under one category, e.g., under Knightian uncertainty, we have here
shown the more traditional breakdown. ‘
3And as we shall see below, the "welfare loss" increases as the
square of its greater-than-normal profits--given the elasticity of
18
difficult to separate "contrived" from "natural" scarcities, we shall
make some attempt in this direction. Actually, by taking a reasonable
choice of periods to investigate we can eliminate many of the above
causes of surplus not attributable to monopoly power. By finding a
long-term average profit ratg, we can expect to eliminate windfalls
-,_____’,,LL
which are, by definition, sporadic or intermittent. 'A weighted
average profit rate for all firms in the economy or in the industry
should, under certain assumptions, also eliminate risk as an explana-
tion of group-average "excess" profits (losers offsetting winners
giving a zero net return). For not all firms in the economy or industry
can earn "excess" profits which can be described as risk rewards--the
existence of risk being rewarded should be proved by losses to other,
less successful firms. If all firms persistently earn 10 per cent
p. a. "excess" profits, it is difficult to describe these earnings as
risk rewards, or to call the industry a "risky" one for investment.
However, if the economy suffers from risk.aversion, or a "systematic
overestimation of risk", this return may persist in the long run.1 k/X
demand, Harberger, op. cit., p. 85. For an excellent short defense of
identifying "excess" profits with monopoly profit see Joe S. Bain,
"The Profit Rate as a Measure of Monopoly Power,"‘gparterly Journal of
Economics, Vol. 55 (1940-1941). Any subsequent references to Bain,
unless otherwise indicated, will be to his book, Industrial Organiza-
tion, and not to this article. We should also mention that Bain further
found that profit rates did not vary continously with the degree of
concentration, although he did, in general, confirm the monopoly-
competition distinction in his "Relation of Profit Rate to Industry
Concentration,"_Qu§rterly Journal of Economics, LXV (August, 1951), pp.
313-314. This position was confirmed by D. Schwartzman, "The Effect
of Monopoly on Price," pp. cit,, pp. 360-361.
1George J. Stigler, ggpital and Rates of Return in Manufacturing
Industries (National Bureau of Economic Research, 1963), pp. 62-64,
ound no evidence of a risk premium in manufacturing, although it was
admittedly a restricted investigation.
19
Long-run considerations should also neutralize most of innova- if
tional profits which are presumed to be removed in the long-period
by the march of successive and successful imitators. Similarly, the
returns erroneously attributed to profit which are actually due to
a specific resource, say naturally scarce land, may be eliminated by
taking a longer view. The alleged profit due to innovations and vary-
ing rates of growth (under the Schumpeterian schema) and not to monopoly
restrictions, even if not completely eliminated by taking the long-view, i
does not appear to be a serious problem. For instance, in the U. 8.,
in the 1950's, there was no special tendency for either the more or
less concentrated industries to grow more rapidly.1
Incidentally,
there are monopoly gains that accrue to other factors that should be
adjusted for in estimating total "welfare losses." Monopoly elements
are in rents, royalties, executive compensation, wages, etc. For
instance, it has been suggested that there are wage differentials in
favor of concentrated industries which are a reflection of this con-
centration and not of divergent skills.2
There are, of course, a number of other adjustments which can be 9
made along these lines. However, the important point that we want to I
.stress now is that the possibility of adjustments of the bare profit
rate makes our position that "excess" profits are entirely caused by
monopoly more tenable. Bain goes so far as to say, \/
1Leonard W. Weiss, Eggnomics and American Industry (New York:
John Wiley and Sons, Inc., 1961), pp. 500-504; also see pp. 511-518.
2Stigler, pp. cit., p. 35; Weiss, op. cit., pp. 505-507; Joseph
W. Garbarino, "A Theory of Interindustry Wage Structure Variation,"
lggarterly Journal of Economics, LVIV (May, 1950), 300 ff. Below we shall
cite contrary studies.
20
be reflected in long-term average excess profits of entire indus-
. . . the only sort of excess pniits which might be expected to \f
\/
tries are monopolistic excess profits. All other types of excess !
l
l
profits are likely to occur sporadically and irregularly, or to
be confined to only part of the firms of an industry . . . Chronic
excess profits are at least_prima facie suspect of resulting from ,J
simple monopolistic restriction, and if so are undesirable.
We hope that even a brief and incomplete sketch of the theoretical
and empirical difficulties of isolating monopoly-caused "welfare losses"
as this will give the reader some flavor for the problems we shall be
encountering.2 What we want to do next is discuss the first approxima-
tion "welfare loss" model for evaluating this loss. Also in Chapter II
we shall discuss Harberger's results from the application of this model.
This will be a prelude to our own extended and revised model which in-
corporates Stigler's criticisms as well as other needed modifications.
_r___~
1
Bain, op. cit., pp. 377-378.
2It should be mentioned that within the economics profession, I
quantitative monopoly studies have not been too favorably received.
Even with some very competent studies that have pointed out clearly
the monopoly-competition dichotomy, the fact that it was necessary to
use such "theoretically imperfect instruments as census industry classi-
fications, interindustry comparisons, and accounting profit rates,"
has caused some reluctance by economists. In other words, the squeam-
ishness stems from the fact that (l) the degree of monopoly being
presumed to be high where economies of scale are important, we obtain
a small difference in P and AC at low outputs in monopolistic indus-
tries. The bias is reversed at large outputs; (2) large errors may
result from census industries that are not the same as theoretical
industries so that small monopolistic industries may be submerged in
large, apparently competitive census industries reducing the ob-
served effects. A similar (same direction) bias results from the
division of large competitive industries into small, apparently
monopolistic census industries; (3) monopoly profits may be capital-
ized under various titles; (4) no adjustment for costs which are
really profits may lead to an understatement of profits in competi-
tive industries since concealment of profits is probably more im-
portant in small than in large firms. See David Schwartzman, "The ,
Effect of Monopoly on Price," op. cit., pp. 352-353 including note 7.
21
In the discussion in Chapter III of our model we shall be careful to
spell out our working assumptions.
CHAPTER II
THE FIRST APPROXIMATION "WELFARE LOSS” MDDEL.AND
ITS APPLICATION BY HARBERGER
A. C. Harberger's ingenious attempt to evaluate the social losses
from concentration is simple,yet revealing. However, he did not use a
”complete” model in the sense that it considers only the effect of
"excess" profits and neglects any redistributional effects (alternatively,
we can say he assumes them equal to zero)- It is also limited by the
fact that resource misallocation might arise from causes foreign to the lg
model--"tariffs, excise taxes, subsidies, trade-union practices, and
1 What we shall do in this chapter
the devices of agricultural policy."
is explain‘the first approximation schema for estimating the allocative
loss from monopoly positions and describe the empirical results obtained
by Harberger from its application to American manufacturing from 1924-
1928. Thus, this chapter will serve as more than the traditional "review
of the literature." By surveying the basic theoretical framework here
we may reserve the next chapter for our theoretical and statistical
modifications‘without having to reconstruct the first approximation model.
To estimate the "welfare loss," the Lerner-Bain approach discussed
above is particularly useful. By assuming a constant long-run AC curve,
and a constant MC curve,for both the firm and the industry, difficulty
lHarberger, op. cit., p. 87.
22
23
of obtaining MC figures is circumvented. Under the constant costs
assumption, the ratio "excess" profits to sales exactly coincides with
the otherwise superior Lerner approach. The former ratio now tells us
by what percentage prices are too "high" or "low" compared to the
optimum. In order to compute the numerator of the 83188 apPTORCh: it 13
necessary to find the rate of return on capital (investment) and sub-
tract it from the economy-wide average rate of return and finally
multiply the resulting figure by the absolute capital base. We shall
discuss this in greater detail below and illustrate it with a few
examples. However, before moving on, we want to emphasize that the
assumption of constant costs is a rather important "wedge" to get the
needed information from the accountant's ledgersfl There are other \
important assumptions,e.g., unitary elasticity,'long-run equilibriumf'
etc., that we shall note as we proceed. So that we do not lose the
proper perspective concerning our findings, we shall list all of these
assumptions at the end of the next chapter.
The central argument for what we shall be doing empirically and
what Harberger did may be succinctly summarized as follows:
. . . conjure up an idealized picture of an economy in equil-
ibrium. In this picture all firms are operating on their
long-run cost curves, the cost curves are so defined to yield
1It has been pointed out that the conditions necessary for con-
stant costs may have been realized in many branches of American indus-
try as a result of the development of "an economy of expensive labor
and cheap capital and of industry accustomed to business fluctuations."
To generalize this assumption to the whole economy assumes, not proves,
that monopoly exists everywhere. Rn'MC must be rising in the relevant
range of output if there is optimum utilization of capacity'--m1nimum
Acééthere is if competition is pure. Malchlup, pp;_g££., pp. 514-517.
24
each firm an equal return on its invested capital, and markets
are cleared. I think it is fair to say that this is a picture
of optimal resource allocation. Now, we never see this idyllic
picture in the real world, but if long-run costs are in fact
close to constant and markets are cleared, we can pick out the
places where resources are misallocated by lookingfgt the rates
pfgreturn on capital. Those industries which are returnipg
higher than aveggge rates have too few resources; and those
yielding lower than avergge rates have too many resources. ‘22
get an idea of how big_a shift of resource it would take to \
egualize profit rates in all industriesi we have to know somg- \
thing_§bout the elasticities of demand for the ggods in question.
In Figure 1 [Figure 2] , I illustrate a hypothetical case. The
industry in question is earning 20 per cent on a capital of 10
million dollars, while the average return to capital is only 10 t
per cent. We therefore build a 10 per cent return into the cost f
curve, which leaves the industry with 1 million in excess profits.
If the elasticity of demand for the industry's product is unity, ;'
it will take a shift of 1 million in resources in order to '
expand supply enough to wipe out the excess profit.1
Price,
Cost
_ "Welfare Loss"
Excess ‘:;><:::’.
...Prnfits, Unit Cost (Inc. 10% on
Incre- \ Capital)
ental
esource
Demand
Quantity
Figure 2.--Harberger's "Welfare Loss" Diagram
1Harberger, op. cit., pp. 77-78 (italics supplied). For the
reader who is accustomed to thinking in algebraic terms, Stigler re-
. stated the above argument in "primitive symbols" in op. cit., pp. 34-35
as follows: "Cost of production per unit are a of labor and ic of capi-
tal (where i is the competitive interest rate), and a and c are constants
if the industry has constant costs. The demand function has unitary
elasticity, so pq a S (where S is sales). The monopolist obtains a rate
of return of mi on his investment. Then if the price is to fall to the
competitive level, output will raise in theratio Sfla-i-ic) - S[(a+mic)
S (a+mic) '
This expression simplifies to (m-l)/Rc [apparently a typographical
25
The first thing to be noted in connection with this theoretical
approach is the pivotal position played by constant costs. If we had
the, perhaps, more typical situation of rising MC, we would get the
following result.
Price,‘ MC
Cost P,AR \ :5
\\~ AC
AC ...K...
MC AR
0 Quantity
Q
Figure 3.--Profit Rates under Rising Versus Constant Costs
' Now profit data [profit per unit being P,(AR)-(AC)] would not give us
the informationwe desire on the difference between P and MC--our
measure of the "welfare loss." It should be observed that, if costs
in American industry are increasing rather than constant, less real- /
location of resources would be necessary to equalize profit rates. \4
This means that the assumption of constant costs, probably, overstates
the "welfare loss" due to monopoly. This is illustrated in Figure 4,
where the "welfare loss" from the constant cost case, MCo(=ACo), is
the whole area in the triangle ABC. Under conditions of rising MC,
MCI, it would be the shaded triangle ADB.
'It should also be noted that it is something of a simplification
to regard the average profit rate as the competitiyg‘rate. Surely,
W” “W“... .
error made this (m-1)Rc in the J.P.E.] , where Re is the ratio of all
competitive costs to competitive capital costs (that is,
Rc = [(a+ic)/ic] ."
U
i
26
Price,
Cost
ACo = MCo
ARC”1
o ‘ Quantity
Figure 4.--"Welfare Losses" under Rising Cost Conditions
this need not be the case,e.g., in an industry with one competitive 3
firm and the rest monopolistic, the average rate would undoubtedly b I;
much greater than the competitive rate. In most cases, this should 3
not lead to any large-order errors.
Also, the portrait of an economy tending toward equality of
profit rates is subject to several qualifications:1 (1) In the short
run, imperfect knowledge of returns on alternative investments or a
lack of desire for profits would cause some dispersion among profit
rates; similarly, unexpected developments and events which call for
transfers of resources requiring considerable time to be completed
would lead to dispersion, but presumably would be eliminated in
rthe long run as knowledge is transmitted. (2) Persistent,long-run
differentials can be anticipated if there are differences among
industries in monetary and nonmonetary supplements to the average
L
1Stigler, Capital, op. cit., Chapter 3, pp. 54-71.
27
rate of return,e.g., the reputed psychic income associated with the
"good life" of farming, as well as teaching, would lead to below average
1 and tax considerations may require higher
returns while risk premiums
than average returns. (3) Finally, in any empirical study, the differ-
ence between the income concepts used in compiling the data and the
income concepts relevant to the allocation of resources is a third
source of dispersion-~perhaps the three most important defects being:
(a) the concept of income appropriate to resource allocation differs
markedly from the notion underlying income tax data; (b) there may be
"excess" salary withdrawals made by the officers of small corporations
who own most of the stock and have a great deal of discretion in taking
income as salaries or as returns on capitel--the difficulty being
that only the second-mentioned form of income goes into our rates of
return;2 (c) the asset values used in computing rates of return have
not been adjusted for price changes.
Perhaps a word of explanation would be helpful on how the quantity
of resources that must be shifted in a function of the elasticity of
demand. More precisely, the value of the misallocated resources is
equal to "excess" profits times the estimated arc elasticity of demand
1If lenders accurately estimate future risks on average, it is
reasonable that they demand a nominally higher rate when assuming larger
risks. However, if the rate is higher only by the actuarial value of
future risks, we should maintain there is no risk aversion.
chConnell's procedure for this contention rested upon the assump-
tion that differing marginal productivities of capital (mpk) among
companies of a given size explain any difference in income. It seems
more reasonable to assume mpk's are the same and entrepreneurial
skills differ. See ibid., pp. 125-127.
w—quA-uw
28
between the monopoly and competitive positions on the demand curve.
If you express the amount of "excess" profit, $1 million in the above
example, as a per cent of sales in the industry, $10 million here,
we would obtain the percentage that price in that industry is too
:
g
"high" ("low" if the rate is less than average) compared to the ideal l
allocation of resources, 10 per cent in this case. Since this ratio
effectively measures the ratio of average price to AC, and MC under
our assumption, a 10 per cent "excess" profit
rate on sales indicates ?
that average costs are 90 per cent of the average price at which
sales were made.
The above short discussion gives us a general picture of what
has to be done empirically. It is then access
that, at least, roughlymeets two conditions.
ary to find a period
First, it is desirable \
to find a period approximating "long-run equilibrium" with no drastic
shifts of demand or economic structure in process. Otherwise, we
could get cases such as an increase in demand
for farm products
(agriculture reputedly being our least monopolistically-dominated
sector) leading to a short-run rise of returns
productive resources flowed into the industry
course, this higher return is not due to monop
on capital until new
in the long-run. Of
oly power but is merely
a high economic rent due to accounting procedures and the natural
scarcity of land. However, by taking long-term profit rates, the
sporadic and irregular components, risk, uncer
tainty, perhaps inno-
vations, would be mainly removed leaving "excess" profits due only to
monopolistic restriction. Of course, this is
want to separate to estimate "welfare losses."
just the element we
Secondly, a period for
3
.
\
\
5
I
i
29
which accounting values were near actual values is desirable. We
know that the accounting profit is biased upward if the price level a
has been rising, and downward if it has been falling. This follows 1
from the fact that accountants typically "measure in terms of dollars
of different purchasing power." They measure current revenues and
costs in current dollars and past costs and investments in past
dollars. In other words, they do not make price level adjustments
in stating dollar values.1
Harberger took the 1924-28 period as a reasonable approximation
to the above mentioned conditions. This period had the additional
advantage of being able to employ Professor Ralph C. Epstein's fine
1For a numerical example of this see Bain, op. cit., pp. 380-
381. If a researcher is not able to find such an ideal period,
adjustments can be made. Although Stigler, Capital, op. cit., pp.
34-37, 49-53, shows, by rank correlation analysis, deflation of book
values to get "real“ assets did not change things significantly; in
fact, the rate of return (after taxes) on all manufacturing was 7.2
per cent a year in both current and stable prices from 1938-1956.
Bain, op. cit., pp. 381-382, using Statistics of Income data, which
we shall also rely on so heavily later, feels although 1936-1940 all-
corporations profit rate as a percentage on equity (after income tax)
can be accepted more or less at face value; the 1949-1953 period,
with a rather sharp price inflation over the preceding eight or
nine years along with its own slower inflation during the period
itself, should include a reduction of one to two percentage points
in the stated rates.
Part of the spurious profit of the above sort is caught in
the Department of Commerce figures in the "corporate profit before
taxes ('adjusted')" account. The "adjusted" refers to an inventory
valuation adjustment. Thus, in 1959 rising prices led to a $0.5
billion deduction (-$).5 billion) from the reported corporate profits
of $47.1 (adjusted a $46.6) billion while in 1953 falling prices
necessitated a $1.0 billion addition +$l.0) to reported profits
(similar adjustments being made for unincorporated incomes). A
similar difficulty, which we shall not be able to do much about, is
the fact that the value of the common stock will capitalize "excess"
profits so as to leave a yield that is apparently "normal." Adjust-
ment of income data will, at least, partially make up for this.
30
work, Industrial Profits in the United States.1
Then, to approximate
a "long-period equilibrium," void of factors causing short-rate vari-
ations, the industry profit rates for the five-year period, 1924-1928,
were averaged. The computed differences among these profit rates, as
between industries, gave a broad indication of the extent of the
resource misallocation in American manufacturing in the late twenties.2
1(National Bureau of Economic Research, 1934). In this book,
Epstein gives rates of total profit to total capital for seventy-three
manufacturing industries. He defines total capital (pp. 595-596) as
capitalization (invested capital of a corporation as measured by the
sum of its preferred stock, common stock, surplus and undistributed
profits with special reserves in most instances excluded), plus funded
debt (capital borrowed from the general public and lending institutions
through the sales of bonds, debentures, notes and other forms of
indebtedness). In general terms, total capital = book capital + bonded
indebtedness. Total profit refers to net income (net earnings after
all business expenses and fixed charges including interest payments on
funded debt have been deducted), plus interest payments on funded debt.
Again, in general terms, total profit = book profit + interest indebt-
edness. The reason that the returns are computed to include funded
debt is that these borrowed dollars perform much the same economic
function as invested capital. If we add interest to the earnings and
funded debt to the capital base, the profit rate on all capital employed
will be lower for most companies since earnings usually exceed the
interest rate charged to the firm, according to Claude Robinson,
ypderstanding Profits (Princeton, New Jersey: D. Van Nostrand Company,
Inc., 1961), p. 73. Another reason for preferring this computation
over the return on equity is that this latter figure "might be quite
different for two economically identical firms, depending on how they
were financed. The firm with the larger debt outstanding will show
higher earnings on equity so long as its interest charge per dollar
borrowed is less than it earns on its total assets. Of course, the
debt-financed firm also is in greater danger of turning in a loss in
bad years since the interest has to be paid regardless of what the firm
earns. In other words, the profits on equity will fluctuate more
widely from year to year for the company with large debts, even if the
economic performance of the two firms is the same." The quoted writer
qualifies his endbrsement of our approach when he says, ". . . when
public utilities are discussed . . . the return on total assets cannot
be compared very easily with those of other industries. At any rate,
it is the return on‘owners' equity that businessmen presumably are
trying to maximize." Weiss, op. cit., p. 144 including note *.
2Of course, a lack of desire for profits or a lack of knowledge of
returns in alternative ventures, etc., could render any tendency toward
31
To better understand the theoretical approach that is employed
in calculating the "welfare losses" by this first approximation model,
we want to carry through the calculations for a particular industry.
Rather than use a hypothetical example we shall use an industry that
Harberger utilized in his estimates. In the process of obtaining the
final results we shall want to elaborate upon some theoretical points
that we touched upon before. The interested reader may then verify the
result by consulting Harberger's tabled estimates.
If the bakery products industry was earning 17.5% return on its.
total capital, it would be earning more than the overall average rate a
for all industries of 10.4%. In order to obtain the absolute amount 3
of ”excess" profits in this industry we would multiply the above '
profit rate differential of 7.1% times the capital base of $242.62'
(capitalization = $236.00, funded debt = $6.62) million. This gives
"excess" profits of approximately $17 million. We then can express
"excess" profits as a per cent of sales to determine by what percentage
the price diverges from the optimum. Since (P-AC)/P = (P-MC)/P = 1-kP,
under our constant cost assumption, the ratio of profits to sales will
. . 4"‘H‘W'I14
”" “
give the desired information as to how "high" or "low" prices are. L/’
‘ ’ ..‘A—a...—_.__ .
3. _.
«...—wan“ » ..
...-“1er
To determine how'much of a reallocation of resources from high j ’{£¢: A?"
,
a
profit to low profit industries would be necessary to eliminate the L
observed divergences in profit rates, it is necessary to know something
about the industry demand elasticities. This may be illustrated in
Figure 5.
equality of rates negligible. But, persistently high profits indicate
the industry is not competitive.
32
Price,
Cost D2
'Welfare Loss"
Excess Profits
G ""'— CIIMC
C
Incremental \\\
Resources \
D2
0
Q R S T Quantity
Figure 5.--Resource Transfers and "Welfare Losses" as a Function
of Elasticity
This diagram illustrates that both the required amount of
resource transfer and the "welfare losses" rise as the elasticity
gets larger. As compared with the resource reallocation indicated
by the rectangle BCQS, resulting with the original demand function Do,
the reallocation rectangle grows to BFTQ with the demand function D1
with the higher elasticity and falls to BQER under the smaller elasti-
city associated with demand function D2.1 Similarly, the "welfare
losses" increase from triangle ABC to triangle ABP for higher elasticity
1In other words, the extent of the misallocation is the value
of the resources that must be brought into the industry for the price
to fall to the competitive level. This value is the competitive
price (0C), times the difference between competitive and monopolistic
outputs,IJ, or the rectangle DGIJ in Figure l. Elasticity comes
into the picture by affecting the size of IJ.
33
and fall to triangle ABE for the smaller elasticity case.1 Thus, we
can see that elasticity is quite important in our study. Harberger
felt that unity elasticity was a reasonable assumption since the
analysis involves the substitution of one great aggregate of products
yielding high rates of return for another yielding low rates and not
the substitution of one industry's products against all other products.
Since we shall return to this point later, we shall only mention here
that we think industry-by-industry estimates of elasticity are more
realistic for determining the relevant magnitudes involved in this
type of analysis.
This brings us to the question of what do we mean by "resources"
when we talk about transferring resources?
. . . resources here . . . mean the services of labor and
capital plus the materials bought by.the industry from other
industries. In many ways it seems preferable to define
resources as simply the services of labor and capital. This
could be done by applying to the value added in that industry
the percentage of excess profits to sales. The trouble here
is that adding to the output of industry X calls resources not
only into that industgy but also into the industries the;
supply it. And by the time we take all the increments in value
{edged of all these supplying industries that would be generated
1The "welfare loss" may be thought of as the sum of the
producers' and consumers' surplus which approximately equals
Increase in price x reduction in quantity. If the unit of output
2
is defined so that the competitive price, 0C (again using Figure 1),
is $1.00, the reduction in quantity, JI, equahgbGIJ, and the "welfare
loss" equals Igcrease in price x elasticityp(monppoly profits). To
2
obtain the monopolist's increase in price per unit of output, or
the monopoly effect on price, as a proportion of the competitive
output price, Schwartzman, "The Burden of Monopoly," op. cit., pp.
627-628, uses the formula TR - l (which a E ), where, TR =
TR-E TR-E
total revenue, E = "excess" profits.
34
by the initial increase in output of industry X, we come pretty
close to the incremental value of sales in industry_x. Of course,
the movement to an optimal resource allocation entails some
industries expanding their output, like X, and others, say Y,
contracting their output. If we really traced through the incre-
ments to value added which are required in their supplying indus-
tries, say Z, we would often find that there was some cancellation
of the required changes in the output of 2. Hence by using sales
rather than value added as our measure of resource transfer, we
rather overstate the necessary movement.
Under the unity elasticity assumption, we may add up all the
plus and minus "excess" profits in all industries to estimate the magni-
tude of the ”desired" resource reallocation. In Harberger's case, to
attain equilibrium would require the transfer of roughly $550 million
in resources from low-profit to high-profit industries. Since Epstein's
sample accounts for 45 per cent of sales and capital in manufacturing,
the extrapolated figure becomes $1.2 billion (using 550/45 s X/lOO yields
x & 1.2222 billion). The tentative conclusion is that manufacturing
misallocation in 1924-1928 could have been eliminated by a net transfer
of roughly 4 per cent of the resources in manufacturing or 1% per cent
of the total resources in the economy.
We now want to estimate how much better off people would be if
the desired resource reallocation was effected. To calculate this, we
may use a formula suggested by Hotelling in 1938 for an analagous
problem.2
l
Harberger, op. cit., pp. 80-81 (italics supplied). For a defense
of partial-equilibrium analysis, e. g. ., against the charge of neglecting
the fact that as prices decline in monopolistic industries, the demand
and cost curves may shift, see Schartzman, "The Burden of Monopoly,"
op. cit., p. 630.
2Since Hotelling's formulation is not immediately obvious, we have
reproduced Harberger's note on it, ibid., pp. 81-82, in toto. "Harold
35
Hotelling's original expression for the total "welfare loss,"
8 2 dpi dqi can be obtained by a simple application of the formula for
the area of a right triangle, i.e., the area is equal to one-half the
product of the two legs, A 2 § leg AB x leg BC. we shall show it on a
per unit basis, i.e., 8 dp1 dqi. Since we know from the previous
discussion that the triangle ABC measures the "welfare loss," we can
estimate this loss by the above formula. We get A = (AB) (BC) /2 =
dpi dq1/2, since AB is, in fact, dp and BC is dq. (See Figure 6 on
the following page.)
Hotelling, 'The General Welfare in Relation to Problems of Taxation and
of Railway and Utility Rates,‘ Egonometrica (July, 1938), pp. 242-269.
The applicability of Hotelling's proof to the present problem can be
seen by referring to p. 252 ff. Be there indicates that he hypothe-
cates a transformation locus which is a hyperplane. This is given us
by our assumption of constant costs. He then inquires what will be the
loss in moving from a point Q on the hyperplane, at which the marginal
conditions of competitive equilibrium are met, to a point Q' at which
these conditions of competitive equilibrium are not met. At Q' a non-
optimal set of prices prevails. These are, in our example, actual
prices, while the equilibrium price-vector P is given by costs, defined
to include normal profits. Hotelling's expression for the welfare
loss in shifting from Q to Q' is k iidpi dq1, where p1 and qi are the
price and quantity of the icth commodity. _We obtain this by defining
our units so that the cost of each commodity is $1.00. The equili-
brium quantity of each commodity under the assumption of unit elastici-
ties is then equal to the value of sales of that commodity. If we
call r1, the percentage divergence of actual price from cost, we may
write the total welfare loss due to monopoly as k42.r12 q1 if the
elasticities of demand are unity, and as kér 2 qi k1, if the elastici-
ties of demand are k , In column 4 of Table , I attribute to each
commodity a welfare loss equal to k r 2 qi. This measure of the welfare
loss due to monopoly abstracts from t e distributional considerations.
Essentially it assumes that the marginal utilkfiy of money is the same
for all individuals. Alternatively, it’may be viewed as measuring the
welfare gain which would occur if resources were shifted from producing
Q' to producing Q, and at the same time the necessary fiscal adjust-
ments were made to keep everybody's money income the same."
36
Price, " "
Cost P A Welfare Loss
h Excess Profits dpi.
MC
B dqi
Quantity
Figure 6.--Graphical Derivation of the Hotelling "Welfare Loss"
Formula
From this form, one can get to the alternative formulation,
ki r12 :11 k1. With r1, the percentage divergence of actual price
from cost, the amount of excess profits, the formula becomes
total sales
(r12 qi)/2. By defining units so the cost of each good is $1.00, we
can use sales for quantity figures. We do this manipulation so that
we can compare a $1,000 car with a $10 radio by saying we have 100
units of car and 10 units of radio, at the defined cost of $1.00. We
may prove this in the following way:
(1) dpi = r1 p1 and dqi = r1 k1 qi, i.e., we turned r1 percentages
into absolute figures.
(2) dpi dqi = (r1 P1) (I1 91) (R1)
(3) up, dq1 = r12 (pq) k1 = r12 41 k 1
(4) if the elasticity of demand is unity, this becomes r12 qi.1
1In other words, when prices are equal to unity then dpilpi =
dpill = r1 is equal to the percentage change from that price as a result
of the deviation of profits from their normal level. Now the percentage
change in quantity will be r1 X k1, where k1 is the price elasticity of
37
Unfortunately, Hotelling's formula is not quite accurate.
His general formula would be strictly applicable here if all
our industries were producing products for direct consumption.
The question thus arises, how to treat industries producing
intermediate products. If we neglect them altogether, we
would be overlooking the fact that their resource shifts and
price changes do ultimately change the prices and amounts of
consumer goods. If, on the other hand, we pretend that these
intermediate industries face the consumer directly and thus
directly affect consumer welfare, we neglect the fact that
some of the resource shifts in the intermediate sector will
have opposing influences on the prices and quantities of con-
sumer goods. Obviously, this second possibility is the safer
of the two, in the sense that it can only overestimate, not under-
estimate, the improvement in welfare that will take place.
We can, therefore,follow this course in applying the Hotelling
formula to our data.
Returning to our bakery products example, we take the $17 million of
of absolute "excess" profits we previously found and divide it by
industry sales of $320 million to get an r1 of 5.31251. Substituting
into the “welfare loss" formula of (r12 q1)/2 yields [(053125)2 x
320] /2 = (.002822) x 160.23.4515616. Less rounding error is involved
if r 2 = (amount 0f "excess" profi£§)2
1 (industry sales)2
L17)2(320) ..ng = 289 ~ 04512.2
(320)2(2) (320)(2) 640 ‘
is used. This gives
~
demand for the ith product. In the case of unity elasticity
a . — IC =
dq1 / qi r1 Since 1. 1 this means dqi /dp1; in general, dqi/qi =
41 Pi
k1 dpilpi or since dpi /p1 = r1 , = k1 r1. To get the absolute change
in quantity, i.e., dqi, we must multiply the percentage change in quan-
tity times the absolute quantity. Thus, riqiki is equal to the absolute
change in quantity. Since the absolute change in price, starting from
an initial situation where the prise is equal to one, is r1,we have an
expression for dpi dqi, namely, r1 qi k1.
1
”Harberger, op. cit., pp. 82-83.
2To review our understanding of exactly how to compute these
"welfare losses" as well as to gain some fresh insight into more of the
economics--at the expense of computational efficiency--of the problem,
38
Using this same basic "welfare loss" formula, Harberger found
the total ”welfare losses" over all manufacturing industries to be $59
million, $26.5 million unadjusted for sample size, or $225 million in 1953
present value terms. In other words, his estimate of the aggregate
loss amounted to less than one-tenth of one per cent of national income or
$1.50 per person in the United States in the 1924-1928 period.
The above discussion covers the main arsenal of this type of
attack on the efficiency problem, as well as some of Harberger's
results from its application. However, there may be flaws in the data
Operating to make any estimate too low-~remember the constant cost
assumption works in the other direction, if there is increasing costs.
For instance, intangibles, such as goodwill and patents, by being
assigned a book value may capitalize monopoly profits. The reported
we want to repeat Stigler's explanation. He explains the theory
through an example of the toilet preparations industry as con-
tained in Harberger's estimates.
”In Epstein's sample this industry earned an average of 30.4
per cent on capital in 1924-28, while the 'competitive' rate
(that is, the average rate in all manufacturing) was 10.4 per
cent. Hence monopoly profits were 20.0 per cent of capital,
and, since capital was $16 million in 1928, monopoly profits
were 0.20 x $16 million = $3.2 million. The competitive costs
of the industry's output were therefore its $20 million sales
minus $3.2 million, or $16.8 million, and we may choose such a
unit of output that the industry was producing 16,800,000 units
at a cost of $1.00 each. The monopoly price of these units
was $20,000,000/l6,800,000 a $1.19. With competition, the
output would be 20,000,000 units and the price $1.00. Since
the loss of welfare due to monopoly is taken as
Ipcrease in output x reduction in price, we may substitute
2 .
our numbers,
3 200 000 x .19
2 a $304,000."
39
profit rate thus understates the actual profit on real capital.1 \V/
Of course, even the elimination of intangibles is not enough for
monopoly profits can be capitalized under many asset titles;J For
example, Weston found mergers and acquisitions accounted for one-
fourth of the total growth of assets of seventy-odd Corporations in
the last half-century. Harberger, for one, discounts this factor on
the grounds that any over-valuation would be off the books by the 1924-
1928 period as much of Weston's merger growth occurred right after the
turn of the century.
1Epstein investigated this somewhat and found excluding intangi- V/ .
bles from the total capital bases made a significant difference in the
earnings rates in only eight of the seventy-three industries. Recom-
'puting the figures for these industries'changes Harberger's estimated
amount of resource transfer from 1% per cent to 1 3/4 per cent of the
national income and changes the welfare loss to $81 million (just over
a tenth of one per cent of national income) To illustrate how this
adjustment was carried out let us take the toilet preparations industry--
again a real rather*than hypothetical example which was taken from
Harberger's study.* We may find the amount of "excess" profits by sub-
tracting the new adjusted profit rate from the old overall rate of
10.4%. Note, although the average would now be higher with the new
higher adjusted profit rates figured in, it is not so computed. This
is because you want not only the'relative" "welfare losses," i.e.,
divergent profit rates among industries, but also the ”absolute"
"welfare loss," figured as the divergence of reported profit rates,
with intangibles in the base spuriously lowering profit rates, from the
"ideal" profit rate excluding intangibles. In other words, if we figure
a new average profit rate excluding intangibles, and it was, say 11 per
cent, this would catch only the "welfare losses" from among industry's
profit rate divergences. We would also like to catch the "absolute"
"welfare loss" by keeping the average at 10.4% for the economy, but
allowing the industries to use new higher profit figures.' It should also
be noted that the new higher industry profit rates are multiplied by the
new lower capital base. However, the adjusted amount of "excess" profits
can not be smaller than before the adjustment.
A simple proof for this can be formulated as follows: Let K =
original capital base, I = intangibles which are‘z 0, TP = total profit,
and 10.4 be the average profit rate before and after the transformation.
The old profit rate = TP/K, the new profit rate = TP/(K-I), the old amount
"excess" profits s (TP/K - 10.4) K which can be written (TP)K - 10.4 K =
a K
TP - 10.4 K ='nf, the new amount of "excess" profits = (IE - 10.4)(K'I) =
TP'1004 (K-I) = 77". Since X Z K'I’fi‘ 7r: K-I QoEoDo
40
Another difficulty with the data is that frequent discounter of
economic analysis: overaggregation. Too broad an industrial classi-
fication makes our assumed high substitutability among the products
produced by different firms within any industry and relatively low
substitutability among the products of different industries less
tenable. The trouble is that in some industrial classifications (e.g.,
Epstein's) remote substitutes produced by quite distinct groups of
firms are lumped together,i.e., the industries are aggregates of sub-
industries. Since it is more appropriate to deal with the subindus-
tries directly, the use of aggregates biases estimated "welfare loss"
downward; but, probably, this error is slight.1
1Ibid., p. 84. "The extent of the bias is proportional to the
difference between the average of the squares of a set of numbers and
the square of the average, the numbers in question being the rates of
excess profit in the subindustries. Consider an industry composed of
three subindustries, each of equal weight. Assume, for an extreme
example, that the rates of excess profit (excess profit expressed as
a per cent of sales) are 10 per cent, 20 per cent, and 30 per cent in
the three subindustries. The average rate of excess profit of the
aggregate industry would then be 20 per cent, and, by our procedure,
the estimate of the welfare loss due to that industry would be 2 per
cent of the sales. If we had been able to deal with the hypothetical
subindustry directly, we would have estimated the welfare loss
associated with them at 2 1/3 per cent of the aggregate sales." He
obtains these figures in the following manner:
Using k r12 qi, & (.20) = k .04 =.02 or 2% for the aggregate
industry; separately, (where qi = sales) it would be:
a(.10)2 = a .01 = .005 '
§(.20)2 e g .04 a .02
a(.30)2 a g .09 = .045
T°“1 '070 4%19-= .02333 or 2 1/31 for
subindustries if computed
directly.
Epstein's data are further complicated by the fact his sample had an
average profit rate of 10.4 per cent while manufacturing activity as a
whole had one of 8 per cent. A correct weighting procedure would adjust
for this apparent overweighting of high profit industries by raising the
estimated "welfare cost" by no more than 10 million dollars. However,
41
The analysis is still not complete in that there may be extra-
monopolistic misallocations arising out of the dynamics of growth,
development, etc., that are disentangled with the monopolistic misal-
locations. Although this is not the sort of thing that one can do
with any great precision, Harberger trys to get his estimate a little
closer to reality on the basis of two props: (1) it is reasonable to
equate monopoly profits with high rates of profit; (2) no more than a
third of manufacturing profits should be monopoly profits. Since
capital is a highly productive resource, he feels this second premise
requires little defense. The first premise is justified on the basis
of (a) observation of the high-profit industries--cosmetics,,drugs,
soaps, autoes, cereals, etc.; (b) the fact given the elasticity of
demand for an industry's product, the "welfare loss" increases with
the square of its greater-than-normal profits--he feels this is an
even stronger reason than (a). Thus, granted (2) the biggest "welfare"
effect is obtained by distributing this monopoly profit first to the
highest profit industries, then to the next highest, and so on.1
this estimate neglects part of the overweighting and this results in
an overstatement of the actual amount of the "welfare loss." This
brings his losses to $2.00 per head (utmore than a tenth of one per
cent of national income.
1In other words, the idea is this. Suppose we say that we have .
a certain amount of monopoly profits, but do not tell in what industries
those profits belong. We can make the "welfare costs" associated with
monopoly very low by spreading these monopoly profits over all indus-
tries, and making the "degree of monopoly" the same in every industry.
In fact, if we were able to do this for all the economy, are could make
the "welfare costs" equal to zero. Different ways of distributing the
monopoly profits obviously will lead to different measures of the
"welfare costs." We get the biggest "welfare cost" by putting the
monopoly profits all in one place, that is, making the degree of dis-
tortion very high in one single area (for remember, given elasticity,
42
After this is done, he concludes the present value "welfare loss”
is no more than a thirteenth of a per cent of the national income
or $1.40 per capita.
Finally, another limitation of the analysis is in neglecting ~//
selling costs, especially advertising expenditures. The difficulty
is that accounts call these expenditures arts, while to an economist;
they are a type of "quasi-monopoly profit," i.e., in the perfectly \
competitive world these expenditures are zero. One way to allow
for this is to make the strong assumption that all advertising is
.persuasive (manipulative) and none informational. Since there is
no way to separate these diverse expenditures, given data in their
present form, this assumption is probably the most useful empirically.
Although Harberger did not make any systematic industry-by-industry
examination of these expenses, he utilized the fact these disburse-
ments ran a little under 2 per cent of sales for his industries.
Even allowing for the maximal distorting effect makes only a slight
the welfare loss increases with the square of its greater-than-normal
profits). But we have to be consistent with the facts and to be con-
sistent.with the facts we cannot assign as monopoly profits any more
than the difference between actual observed profits and the normal
profits on the invested capital in that place. We distribute our
given amount of monopoly profits first to that area where the diver-
gence is greatest, than to where it is next greatest, and so on. In
this way, for any given amount of monopoly profits that we want to
distribute,we are getting the biggest "welfare cost" that we can,
consistently with the observed data.
43
difference raising his estimate of the "welfare cost" to $1.50
per person.1
This completes our discussion of the first approximation model
for estimating "welfare losses." In an actual application of this
type of model, Harberger found the estimated "welfare losses" in our
economy (assumed to be entirely composed of manufacturing) to be
quite small. moreover, Harberger felt his treatment of intermediate
products, his assumption of constant costs and unity elasticity, and
his attributing to monopoly an implausibly large share, one-third,
of manufacturing profits, all tended to overstate the "welfare
losses!” Therefore, he was quite surprised to find that the total
1Ibid., p. 86, note 6, is again worth footnote space. " . . .
1; should be pointed out,also, that the_general conclusions reached
i3 this paper are not closely dependent on the precise data used.
Suppose, for example, that we had observed the following situation:
industries accounting for half the output of American manufacturing
were charging prices which yielded them a 10 per cent 'monopoly
profit' on sales while the remainder of industries earned a con-
stant rate ofreturn of profit on capital (here called normal profit)
but no more. If we were, in this situation, to reallocate resources
so as to equalize profit rates in all industries, the prices of com-
petitive products would rise and those of monopolistic products
would fall. If demand for the product of each sector were assumed
to be of unit elasticity, we would estimate the gain in welfare ‘
incident upon the reallocation of resources at .125 per cent of total
industrial sales. This would be just about a tenth of a per cent of
the national income if the ratio of manufacturing sales to national
:income approximated the 1924-28 figure. The estimated welfare gain
is obtained as follows: Under our elasticity assumption, prices
would rise by 5 per cent in the competitive sector and fall by 5 per
cent in the monopolistic sector, and quantities would change inversely
by an equal percentage. Taking 100 as the aggregate sales of manu-
facturing, the change in output in each sector will be 2.5, and
taking 1 as the index of initial prices in each sector, the change in
price in each sector will be .05. According to the Hotelling formula,
the welfare gain coming from each sector will be k (2.5) (.05) and
when these gains are added together the aggregate again turns out to
be .125." (Italics supplied.)
44
figure was less than a tenth of a per cent of the national income.
To be sure, he recognized this is not a trival figure--over $300
million--especia11y in light of neglect of redistributional effects,
other malallocative effects,.other selling costs, etc. However, his
final conclusion tended to minimize the importance of the monopoly
element in the American economy.
Our economy emphatically does not seem to be monopoly
capitalism in big red letters. We can neglect monopoly
elements and still gain a very good understanding of how
our resources are allocated. When we are interested in the
big picture of our manufacturing economy, we need not
apologize for treating it as competitive, for in fact it is
awfully close to being so. On the other hand, when we are
interested in the doings of particular industries, it may
often be wise to take monopoly elements into account. Even
though monopoly elements in cosmetics are a drop in the
bucket in the big picture of American manufacturing, they
still mean a lot when we are studying the behavior of this
particular industry.1
1Ibid., p. 87. Cf. Bain, op, cit., p. 384, who, while
admitting the 2, 3, or 4 per cent share of the national income going
to "excess" profits may be small and that their total elimination
would not change the national distribution of income or average rela-
tion of P to AC( =:MC) much, is inclined to regard these "excess"
profits as important for their micro significance despite their rela-
tive aggregative unimportance.
CHAPTER III
MODIFICATIONS ON THE FIRST APPROXIMATION MODEL
To any serious and objective researcher of monopoly problems,
the dangers of attempting to wring economic information out of raw
accounting data are obvious. This is especially unfortunate if one
is attempting to measure monopoly power by "excess" profit rates, as
we are. One scholar has gone so far as to remark:
The unadjusted accounting rate of profit, as computed by the
usual methods from balance sheets and income statements, is
ptima facie ad absolutely unreliable indicator of the
presence or absence either of monopoly or excess profits in
the sense defined. The relationship between price and account-
ing average cost tells us nothing about the degree of monopely power
and little about the extent of excess profits. . . . If
accounting profit rates are unreliable as absolute magnitudesi
they should be even less reliable for purposes of comparison.
Fortunately, there is a way out of this academic dilemma--
adjustment of the data.2 In Chapter II, we investigated Harberger's
efforts in this direction. We feel, for the most part, he did an
1Joe S. Bain, "The Profit Rate as a Measure of monopoly Power," \//
‘Qparterly Journal of Economics, Vol. 55 (1940-1941), pp. 291-292.
2Even ibid., p. 292, is willing to admit "As unadjusted
accounting rates are unreliable for our purposes, so a proper scheme
of adjustment of accounting data may provide an approximate measure
of monopoly profits. From any set of accounting data it is concep-
tually possible to compute a theoretical profit rate of the sort
defined above, and is a feasible statistical task actually to produce
a fair approximation to such a rate."
45
46
excellent job in eliciting the desired kind of economic information.
Unfortunately, we are not completely satisfied with the approach.
This is unfortunate for, if we were, our job would be merely one of
bringing more recent and extensive data on the topic. In this chapter,
we shall describe the variants we shall make on the basic model.
Particularly important in this connection are the fruitful avenues
suggested by Stigler in his review article.
We shall be modifying the accountant's data considerably in our
estimation of the resource misallocation attributable to monopoly
elements in our company. For in the estimation of "welfare losses,"
we are interested only in counting that quantity of assets that would
be held by purely competitive firms in "long-run equilibrium. "
, .. -—~...i.
6—- ..,.
IObviously, then, we want to exclude such intangibles as patents, trade- :-
marks, franchises, goodwill, etc., from our capital base or we would,
in effect, be capitalizing monopoly profits. It is only fair to mention
that the statistical modification we shall be making from Harberger's
approach are not a result of his neglect, but, because the data were not
available to him in the form needed. For instance, detailed information
on advertising expenditures is now available on an industry-by-industry
basis. He used a figure estimated for all of manufacturing and not for
specific industries (2 per cent of sales). On the other hand, some of
our other changes will be of a more substantive nature.
To illustrate the general nature of the kind of adjustments that
are necessary in moving from an accounting to a theoretical rate of profit
we have included a rather lengthy passage from Bain. This can serve as
a "jumping-off point" into this difficult terrain. Bain's comments should
47
also serve to heighten our admiration for Harberger's study-~for he has,
through one avenue or another, covered many of these points in his
remarkably brief paper. The portions not so covered are, with our
presently imperfect data, still in the "unreachable" stage of economic
analysis or require such an intimate knowledge of every American indus-
try as to be virtually impossible to any single researcher.
On the one hand, it is necessary to examine the annual net
income figure (inclusive of interest) shown by the accounts,
and to ascertain from an examination of past records any im-
portant over-or-understatement of theoretical costs resulting
from the original valuation or the method of revaluation of
depreciable or depletable assets. The performance of this
task seems to imply a general examination of the conditions
of acquisition of important blocks of assets, a thorough under-
standing of the operations of the firm, and an appraisal of
the current competitive valuation of assets in use. Cognizance
should also be taken of the apparent affect of arbitrary anti-
cipations of loss in the form of writedowns of assets from time
to time. Particular attention should be given to (l) the rela-
tionship of depreciation charges to the theoretical norm, and
(2) the costing of resources used, to ascertain whether the costs
listed approximate the current competitive rent of these resources.
Such an adjustment procedure could obviously have meaning only
if pursued for a considerable series of consecutive years.
On the other hand, the asset total should be examined to
ascertain what assets are excludable in toto from the theoretical
rate base, what assets are held in amounts in excess of the
theoretical norm, what assets have original valuations which ‘3
seem to include monopoly profits, and what assets have been re-
valued in a manner which understates their probable current com-
petitive value. Intangibles of most kinds, idle land, and
holdings of depletable natural resources, for example, are ex-
cludable in toto from the competitive rate base, the last item
on the condition that currently used resources are entered as
costs at their competitive rents. . . . Original asset valuations
should be closely examined for the possible inclusion of capital-
ized monopoly profits whenever the items involved are included \
in plants of firms acquired in toto by purchase, or through
merger or reorganization, and particularly when large capital 1.
stock rather than small cash transactions have been involved. In
these cases "original cost" is most likely to lose touch with value
in a competitive market, and adjustments are most likely to be re-
quired. . . . A rough check for the presence in the asset total of
48
obviously eliminable items {like long redundant or obsolete
capacity) is possible . . .
Keeping these general suggestions in mind, let us investigate
some more specific modifications we might make on the previously
described model. An excellent place to begin our reformulation is
with the Stiglerian critique mentioned above.2 Two suggestions that
he has made are particularly important. In fact, any possible modi-
fications of the general order of magnitude from Harberger's estima-
tions are likely to stem from these changes. The first of these con-
cerns the scope of his coverage and the second concerns his unitary
1Ibid., pp. 292-293. The difficulty of profit figures is also
discussed in Weiss, op. cit., pp. 144-146, 501-508.
2For the first point see Stigler, op. cit., p. 35. It should be
added that Ruth F. Mack's discussion, op. cit., p. 89, of Harberger's
paper covers some of the same ground. She feels the three most
important doubtful aspects are: "First, the notion that profits are
an adequate measure of monopoly due to maldistribution of capital
has often been called into question. More damaging is the second
problem: neglect of maldistribution of other factors of production
that might be a function of monopoly. Monopoly certainly can yield
inefficient use of labor and materials as well as of capital. This
would mean, in effect, a departure from some proper figure for value
added, or perhaps even total costs, rather than simply for profits.
I ask, in other words, whether the horizontal cost curve to which
Harberger adds the 10 per cent profits may itself be too high, from
the point of view of consumer welfare, because of monOpoly elements
in labor or material costs, because costs are included that consumers
under truly competitive conditions would not elect to pay for (high
marketing, advertising and packaging costs, for example), because of
restrictions on a potential rate of technological change. Finally,
toward what other less than optimal results does monopoly contribute:
maldistribution of income, inflexibility in all sorts of adjustments
including prices to changes in economic conditions--to pick two at
random."
49
elasticity assumption. First of all, the competitive rate of return
on capital should be computed for the entire economy, not just for
the manufacturing sector. The "welfare loss" in manufacturing would
swell if the competitive rate of return were lower. However, since
monopoly is presumably more important in manufacturing than in the
remainder of the economy taken together this would tend to exaggerate
the monopoly loss. The understatement of "welfare losses" can be
interpreted in terms of "absolute" and "relative" "welfare losses." 1
In Harberger's case, if he had used the 6.2 per cent (after the deduc-
._\_ J
tion of Federal taxes) return on capital for all corporations engaged
in manufacturing, trade, finance, and mining, in 1924-1928, found by ~
Epstein from official income tax data,1 instead of the 10:4 per cant
figure for the manufacturing sample, Harberger would not have affected
the "relative" "welfare loss" among industries as all "excess" profit
rates would have been raised from X% - 10.4% to XZ-6.2% or 4.2% (where X
“....
= rate of profit on capital). However, the "absolute" losses would
have gone up fofimanufacturing as a result of this new lower average
profit rate.
The possible overstatement of loss from using manufacturing data
refers to the fact that, since monopoly is presumed more important in
manufacturing, any simple "blow-up" of its loss, say doubling it if
manufacturing accounts for half the sales and assets in the economy,
would surely overstate the case.
\flEpstein, op. cit., pp. 24-25, 49-51.
50
Ideally what we want is profit rate figures for all types of
business establishments, sole (single) proprietorships (SP), partner-
Mwmom -—-—4v-ru-1. raw 11-
ships (P), and corporations (C), for all the various industries. This
we have tried to do for the five-year period, 1956- 1957 period to the
m1-...~.r-_.M...-.
..............
1960- 1961 period. The data for these years were obtained from the
4.“.— W
urea" '
0”
“WW-HM“...
Statistics of Income-~for our purposes, undoubtedly, the best available.1
The period and source were selected for a number of reasons. First of
all, in comparison to earlier years, the data are better and more
reliable as time goes on,i.e., the data for 1960 are superior to the
1950 data, the 1950 are superior to the 1940, etc. It is better than
other data because the IRS gives income statements in some cases and
balance sheets--which allows us to segregate specific accounts, e. .g. .,
“Mt—.83..“
advertising. It is more reliable, as time goes on, with advancing
\\_____,_._.—...
sampling techniques and larger samples available. Secondly, we wanted
years not too near Harberger's period so,after some attempt at standard-
ization of techniques is made, we can get a rough idea if the estimated
"welfare loss" is rising over time. Finally, after the proper adjust- k
ments are made, it is probably as close to a "long-run equilibrium
period" and accounting values are probably as close to actual values as'
I
I
t
I
I
E
l
1
\
any of the intervening years since 1924-1928.2 B
1Stigler, Capital, pp. cit., p. 7, says "Aside from presumably
minor problems of nonreporting and postaudit revisions, this material
is comprehensive in scope, if not always in detail."
2We originally intended to compute profit rates for a longer
period, say 10 years, but, figures showing unadjusted rates for such
length periods convinced us that the results would not differ signifi-
cantly. It is also important to know that the dispersion of profit
rates is relatively greater in years of depression; industries cannot
adjust to sudden decreases in demand as well as they can to increases--
apparently, because fixed capital is easier to increase than to decrease
in the short run. See, Stigler, Capital, pp. cit., p. 6.
51
To be quite honest, the data for these years are not entirely
satisfactory. First, there is no complete income statement and balance
sheet information for all types of establishments for all five years.
All show, at least partial, income statements for most of the years
(P and SP for four of the five years, C for all five); but, only C
have virtually complete balance sheets for the five years. The P only i
have their balance sheets for the 1959-1960 period. Even here only ‘Jj
44.5 per cent of the firms that filed income statements did the same
for balance sheets--though the figure went over 90 per cent in some
particular‘industries. As a result, to use these data we had to blow
them up to represent all P, as well as, assuming this one period was
representative of all four periods. Werst of all, SP only show income
statements. We computed the rates of return on capital for P on the
basis of the "partners capital" account and then assumed rates of i
return for each industry of SP was the same as it was in P. This
allowed us to get back to total capital estimates since we have net
profit figures. However, fragmentary evidence of "excess" salary
withdrawals in small companies warns us that the P account,"partners'
capitalfl may be a bias estimate of the "real" capital investment.1
!
.
Alternatively, we have adopted the procedure used by Stigler to \¢
3
2
fiestimate the capital of noncorporate enterprises-oonce annual data on
ireceipts (sales) are available.2 His estimate of the noncorporate
1E.g., see Joseph L. McConnell, "1942 Corporate Profits by Size
of Firms," Survey of Current Business (January, 1946), p. 11.
2Stigler, Capital, op. cit., pp. 7-8, 114-118, 221.
52
sector is based upon the ratio of capital to receipts in small corps,
.orations (which resemble noncorporate enterprises more closely than
a"
'5.
:n a
they resemble all corporations). It would be undesirable simply to
use the ratio found in the entire corporate sector because: (1)
most noncorporate enterprises are small; (2) small corporations
typically have relatively low ratios of capital to receipts or sales.
The second fact is documented in Table 1, from Stigler, where it is
shown that the ratio of assets to sales is almost twice as large in
the asset class over $100 million as it is in the total asset class
under $50,000--a similar pattern was observed within two-digit
industries. In our estimates, we also tried the $0-25,000 total
asset class ratios. Incidentally, a minor technical departure from
Harberger and Stigler is undertaken when we used the more easily
obtainable business receipts (gross sales plus gross receipts from
operations) in place of sales. The interested reader may find all
the details of the problems we encountered and their proposed resolu-
tion in Appendix A.
We also utilized the balance sheet information of P for estim-
ating intangible assets and royalties. However, since advertising
data are not shown for noncorporate industries, we had to use the
percentages prevailing in C--this will probably lead to an overstate-
ment of the "welfare losses" since in retailing, which is more
important in SP and P than C, wasteful advertising is less significant
than in the manufacturing.1 A further difficulty, which we shall not
l
Weiss, op. cit., p. 511.
53
TABLE 1
RATIO OF ASSETS TO SALES IN MANUFACTURING
CORPORATIONS, 1947
Asset Class Ratio
($000's) Assets to Sales
Under 50 .357
50 - 100 .394
100 - 250 .411
250 - 500 .432
500 -1,000 .447
1,000 -5,000 .508
5,000 -l0,000 .592
10,000 -50,000 .647
50,000 -lO0,000 .642
100,000 - and over .625
All . .625
SOURCE: George J. Stigler, Capital and Rates of Return in Manu- ‘o/"
facturing Industries (National Bureau of Economic Research, 1963),
p. 116. His figures were based on Statistics of Income for 1947.
54
be able to go into because of data difficulties, is the bias resulting
from the fact commodities differ much less with respect to total
selling costs than with respect to advertising expenditures. Unfortun-
ately, we do not have a breakdown of selling expenses other than the
advertising budget. \I
Another difficulty is the problem of comparing the three forms
of enterprises on an industry-by-industry basis. In general, there
are more industry divisions for SP than P which itself has more than C
(there are more service industry classifications for SP than C., 3-8-:)-
This means, for comparison purposes, it is necessary to lump together
various industries. All of which means we are often comparing non-
homogeneous entities among the three types of business enterprises.
Furthermore, the modified SIC classification used by the IRS is so
aggregative that the loss of detail in industries results in the sub-
industry bias mentioned above. Finally, the changes in-the Standard
Industrial Classification make year-to-year comparisons more hazardous
(especially the rather significant changes in 1958-1959).
There are a few other general things which, while making our
analysis more valid, make comparisons with the Harberger-Epstein
findings less reliable. In finding rates of return on capital, in place
of Epstein's ”capitalization" (defined above), which in most cases
excluded special reserves, we have used the roughly analagous concept
of net worth. This latter concept includes preferred and common stock,:
paid-in or capital surplus, surplus reserves, and earned surplus and
undivided profits. Also, in figuring total profit, we have used the
available data in the account "interest paid." Epstein had to adopt
55
a 5% per cent average interest rate since a breakdown was not available
(interest computed as 5% per cent of funded debt was added to income
to determine profit earned on total capital).1 Another change is that
, 4......r
in computing "excess" profit as a per cent of sales, the Lerner-Bain
ll—Q,I.D"V. ...1 .. ._ .... _. ‘¢.L¢_\ -n
”——
index, we used business receipts for sales and did it on the basis
pf the fiye-yesr average for business receipts. Harberger did nOt
average the sales but used the 1928 figure instead--and, of course, used
sales instead of business receipts. There are a number of other
accounting changes we have adopted that we shall discuss below.
We have also computed the rates of return using both an average
total capital base and an average equity base. Our position is that one
..- a... v>*
“
ShOuld be free to choose the approach he feels most appropriate for
the problem being dealt with. Some feel when dealing with problems of
resource misallocation in general, an attempt should be made to get the
total return to capital in an industrial Segment,relative to the total
I
44
amount of capital in that sector. However, when dealing with the prob-<2
lem of monopoly, some feel we should be concerned with the amount of
greater-than-normal profits and to get at this we should look at equity
capital only, determine the amount of greater-than-normal profits and
the percentage that this bears to the value of production.. One can
then use the estimated elasticity of demand to determine the size of
the "welfare cost" associated with the distortion.
In the final analysis, Harberger's restriction to only the manu-
facturing sector may contain even a more fundamental error than
1Epstein, op. cit., p. 601. In "funded debt," we have included
loans from stockholders and both long-term and short-term bonds, notes,
and mortgages .
56
indicated above if viewed in terms of the "second best” conditions.1
These conditions tell us: (1) if the Paretian optimum is unattainable
a "second best" optimum requires a general departure from all the
Paretian optimum conditions; (2) there are unlikely any simple suffi-
cient conditions for an increase in "welfare" when a maximum can not be
obtained. Put more simply, this means if there are a number of exist-
ing divergences, the reduction of one of these-~the others all remaining
constant--will not necessarily lead to an increase in economic welfare
(perhaps even diminishing it). All of this means that "piecemeal welfare
economics" which applies "welfare" rules, which spell Paretian Optimum
if ubiquitous, to only a small part of the economy may move the economy
away from not toward a ”second best" optimum. This means Harberger's
estimation of the "welfare gain” by applying the Lerner-Lange "Rule" to
manufacturing alone may be spurious, i.e., its application may diminish
the general productive efficiency of the economy and the welfare of its
members. This gives us even a more important reason to heed Stigler's
suggestion for a more complete analysis.
The assumption of unity elasticity is also of questionable
validity. Stigler, for one, feels this is an important explanation
of Harberger's low figure for the total "welfare loss."
A monopolist does not operate where his marginal revenue
is zero. A loosely coordinate set of oligopolists might
1See Lipsey and Lancaster, op. cit., as well as the previous
references mentioned above. Incidentally, the appellation "second
best" is derived from the above mentioned fact that the optimum is
achieved subject to the constraint(s) preventing the Pareian optimum.
57
operate where industry marginal revenue is zero, but only
because their monopoly power was very weak-~and it seems un-
desirable to assume that oligopolies are competitive. In any
event, the assumption seems empirically objectionable: most
industries have long-run demand curves Which are elastic.
And in Harberger's model, welfare losses go up when the
elasticity of demand increases.
In order to get a feel for the kind of changes different elas-
ticities would yield, we have computed the "welfare losses" using
elasticities of l and 2.2 Perhaps more realistically, we would like
to estimate the losses on the basis of actual industry-by-industry
elasticity estimates.
Since k1, the price elasticity of product demand, in Hotelling's
formula for measuring "welfare losses," 1; Z r12 q1 k1 (where r1 is
the percentage divergence of price from cost and qi the quantity--all
of the ith commodity) plays a rather pivotal part in our estimates, it
is worth spending a moment on the details of our estimates.
Our first thought of collecting existing elasticity data for
industries was thwarted when we.discovered that most of these data
were in the wrong form--firm instead of industry estimates--or for the
wrong time periods--not for the 1956-1957 to 1960-1961 period--or more
importantly in most cases the data just did not exist in any form.
1Stigler, op. cit., p. 34.
2Schwartzman, "The Burden of Monopoly," op. cit., pp. 628-629,
says that "k [elasticity] is unlikely to have a numerical value
greater than 2" for it "refers to the industry demand curve rather than
to that of the individual firm; the demand elasticity of General Motors
is greater than unity, but that of the entire industry may not be.
Harberger's estimates of resource allocation are for whole industries.
Mbreover, if we are interested in the value of resource misallocation
by monopolistic industries as a group, the relevant demand elasticity
is less than the average of the individual industry demand elasticities."
We did not show these latter estimates since the reader may merely
multiply the first by 2 to obtain it.
58
Since any rigorous, detailed investigation of the relevant elasticities
would be a thesis in itself, we searched for some relatively efficient
but computationally easy estimation procedure. We were fortunate in
finding two methods which roughly satisfied these requirements.
The first of these we shall refer to as the Dorfman-Steiner-
Telser proposition. This proposition states that:
if average variable cost is nearly independent of scale then ‘i
the reciprocal of the advertising intensity is an upper bound
to the price elasticity. Thus, for example, if advertising
outlay is one-half of total sales, the price elasticity at
the optimal output is between one and two. Or, if the adver-
tising intensity is one per cent then the price elasticity is
less than 100. . . . This analysis leads us to predict that
heavily advertised products should exhibit lower price elastic-
ities than little advertised products . . . considering what '
products are heavily advertised lends it credence. Judging
from the §tatistics of Income the most heavily advertised
products are perfumes, cosmetics, other toilet preparations,
drugs, and patent medicines. It seems plausible that the ,.
fi s making these products face demand schedules of rather
low elasticity.1
Unfortunately, the estimates obtained in this manner, while
’perhaps useful for relative dispersions among industry elasticities
finialmost worthless for absolute purposes. The main difficulty is
that the rationale is developed for the firm; but, we must apply
1Lester G. Telser, "How Much Does It Pay Whom to Advertise,"
Proceedings of American Economic Review (May, 1961), pp. 197-199.
It should be noted that ibid., p. 198, says ". . . the advertising
intensity is probably closer to the marginal advertising intensity
assuming increasing average variable cost than assuming constant
average variable cost. Hence the easily measurable number--the
ratio of sales to advertising outlay--may be even closer to the
price elasticity (though it is no longer an upper bound to the elas-
ticity) for increasing than for constant marginal production cost."
See R. Dorfman and P. O. Steiner, "Optimal Advertising and Optimal
Quality," American Economic Review (December, 1954), pp. 826-836,
as well as the first source, for the theoretical defense of this
proposition.
59'
it to the industry. This means the more competitive the industry
(i.e., the less the firm blends into the industry) the less reliable
are our estimates. Thus, in industries such as agriculture we get
relatively elastic industry estimates which in reality should be
firm estimates. However, since the estimates may be useful fog at
least, getting relative relationships, we have included the theore-
tical proof for this proposition (as it is short and straightforward),
as well as the estimates we obtained in Appendix C.
Fortunately, we have another computationally easy and analy-
tically reasonable method of estimating elasticity. This fOrmulation
follows right from the definition of elasticity,i.es,elasticity
B aversge value (A)
(7?) average value (A) - marginal value (M) (similarly, it is true
that A e M Mi: A 711).1 'Analytically we can
__22.__
-1 - :
say that sincefiihe difference between A and H is the force operating
to pull A up or down, we may measure the degree of this force by the
elasticity--a pure number independent of units and dependent on
proportionate and not absolute changes. We may easily prove that
'72 a A/(A-M) or 7’): P /(P-M) in the following manner:
For any demand law p = qu), we may obtain total revenue (R) a
quantity (q) times price (p), i.e., R = qp = q W (q); and average
revenue (AR) 2 (pq)/'q= p, Differentiating R with respect to quantity
gives us a marginal revenue (MR): 2:: sgésgl- = p + q gfi- .
1E.G.,see Joan Robinson, The Economics of Imperfect Competition
(London: Macmillan Company, 1933), p. 36. Since for a rising curve
H J'A, the elasticity of a rising curve is negative here-~which is fine
so long as we are consistent.
60
Substituting these average and marginal values in the purported
elasticity measure gives
- - _ - - £12 .. .. 42.51
nL—A/(A M)-p/(P M>-p/[p-
' £5 3 g ran .. 87.-a m
(1) Unadjusted data E m i C‘ 3 g 8 8-§ g g g 3 3 3 fl z-: 3 g 3 m 8
(2) Data adjusted for intangibles, m m 3 3 5 .4 g 3 % «14 Tm a o m N m m H w o H o 5 u m u H M
-‘ =5 «'40 .1: N-r—l U71: GENO >xlflr-i N—i'U “NT! NH 1:! '0 83>
“”1“": ““d‘dvemmg :22 3.52.: 2.5 5.2 2.2.232: 3°: 3°: 3:: *s :3 52a: 2::
(inpsrenthesis) «iHO o ecu oo.Nm
m va m N a m N ma.Nm - oo.NN
m NNV N N N m N mm.NN - oo.oN
m NNV m m o N m mm.mN - oo.mN
e NNV N e N m o am.NN - ooooN
m Ame o o N o N mm.mN - oo.¢N
m va m N m oN m mm.mN - oo.NN
m Nev N N m m o m¢.NN - oo.oN
NN NeNv NN N NN NN mN mm.m - oo.m
m ANNV NN N m e mN mm.N - oo.o
0 ANN m m N N o am.n - oo.e
N Nev .q n n e m oo.n - oo.N
0 Nov o o N o o m¢.N - oo.o
wEumH
quucH me cowumuomuouuuouwv< womanhood:
Q a . NAOM
.NeN NON
umumsfie<
\/
we coauuzv NN voguuaq
Nm uoauuzv
comm Nmuwmso annoy Ausuhuvuz now uwuuo>< comm muqavm Anus» omen Mundane Hanna
uufiz Hov.owuuo><
NaNNnao mm .m you mauasNNmm
oNNmm Numnm oocmNam m nuNa o chaNaaoo
nNN cocoa: .
uwmonuauumm aw mango momma amuou ooonNaomvmmmHu
uoamm Hmuov oooNom:0m.~suwnmo mm.mm How mommawumm
oNNam Noosm wuanNmm o NNuam nNNa o weNaanoo
ll Jilll
Hmmmv «magnuucuumm oHom can
.Amv unannuoauumm NAUV acouuuuoauou n Amm.NN guesses sags «No2 .N «None mom “weapon
.mousuavaumxm chuNuuo>v< u .>v< can “mowuachom u .hom mmoNanamucH n .ucusx
.mm Nm Now NomNuommN ou mnmNnNmmN voNuud uaoznusom
onu can No you Nomanooma Ou NmmH-ommN mousse umohno>wm on» How mauouou mwmuo>m con: commas
mo¢.oN NNNN.mNV NNN.sN noa.oN mm¢.eN om¢.mN NeN.NN an»:
co co co co oo oo Nance
m Nov n m n N N capo vac oo.Nm
oN NNV a NN 0 ON N ma.Nn - oo.NN
oN ANN N o e N m mm.NN - oo.oN
N ANN a o m o N am.aN - oo.mN
0 ANN m m m .o o NN.NN - oo.oN
o NNNV N N m N m mm.mN - oo.eN
0 Nov NN N m o m .NN.NN - oo.NN
n va m o e N m NN.NN - oo.oN
N Nov e o m o m a¢.m - co.»
m Nnv e m N m e mN.N - oo.o
n Nev m N N m N mm.m - oo.¢
N NNV N N N N e ma.n - oo.N
o Nov o o N o o am.N - oo.o
2v . N3 3 8V 3 as any
fig 350:» N vogue: m woman: . n vogue: 0933033 5
0835.. was sowusuomuoOuouomonuuuoauu265 mo use—52 3am ~3on
TABLE 4
RANK CORRELATION OF UNADJUSTED AND FULLY ADJUSTED PROFIT RATES BY METHODS l, 3, 4, 5, 7, AND 8*
Variables
X(l) = Unadjusted Profit Rate by Method 1
X(2) = Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 1
X(3) = Unadjusted Profit Rate by Method 3
X(4) = Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 3
X(S) = Unadjusted Profit Rate by Method 4
X(6) = Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 4
X(7) : Unadjusted Profit Rate by Method 5
X(8) = Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 5
X(9) = Unadjusted Profit Rate by Method 7
X(lO): Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 7
X(ll): Unadjusted Profit Rate by Method 8
X(lZ): Fully Adjusted (for royalties, intangibles, advertising expenditure), by Method 8
Rank Correlation Coefficient Matrix
Variable. X(l) X(2) X(3) X(4) X(S) X(6) X(7) X(8) X(9) X(lO) X(ll) X(12)
X(l) 1.000000 .912587 .884357 .807724 .898250 .815060 .836899 .789997 .788441 .721145 .784718 .728880
X(2) 1.000000 .817505 .887358 .888914 .932315 .843123 °896138 .801945 .841011 .830897 .859438
X(3). 1.000000 .928091 .860239 .788219 .670909 .647458 .836899 .793832 .658683 .637172
X(4) 1.000000 .839233 .876077 .686913 .750653 .858905 .909808 .697472 .753001
X(S) 1.000000 .941539 .797666 .791442 .800222 .770936 .870019 .845265
X(6) 1.000000 .762378 .838177 .761545 .823340 .855238 .905236
X(7) 1.000000 .954098 .881078 .814226 .944762 .884004
X(8) 1.000000 .856682 .878855 .933537‘ .948310
X(9) 1.000000 .945374 .846457‘ .813250
x(1o) 1.000000 .817338 .862050
x(11) 1.000000 .959426
x(12) 1.000000
SOUchg See Table 2_ *These different profit methods are outlined in Table 3 and discussed in Appendix A.
81
79
position for unadjusted as for fully adjusted estimates, the latter
estimates may be at twice or three times (to take a hypothetical
example) the level of the former. And for many problems, such as
"welfare losses," the absolute levels are quite important. We shall
have occasion to return to this question below after we have actually
shown our estimated "welfare losses."
Before leaving profit rates, there are three interesting rela-
tionships we want first to explore. First of all, in (1.1) we check the
relationship between profit rates on the one hand and intangible
assets, royalties, and'advertising expenditure on the other. Secondly,
in (1.2) the same relationship is examined except we drOp intangibles
for the equation. Finally, in (1.3) we investigate the relationship
of intangibles with royalties and advertising expenses. These are
based on data for corporations for the five-year period, 1956-1957 to
1960-1961.
Our estimating equation are of the following type:
(1.1) ‘771
U _
(1'2) 7”, 13 ‘ a2 + b22 R13 + b32 AEij
where
(103) 11.1: 83 + b23 R11 + 7333 AEij i a 1, o o s , 60 j = 1’ o o s, S
70"11 a the profit rate on average total capital in the ith
industry for the jth year (i.e.,7‘r'1‘1 = net income
after corporation taxes-royalties + interest payments, all
divided by average total capita1-intangible assets»
80
intangible assets in the ith industry for the jth year.
11:
R1:, = royalties in the ith industry for the jth year.
AEij = advertising expenditures in the ith industry for the
jth year.
We than have, for our estimated relationships, five cross-section
equations. If the disturbance terms in a given year are dependent on
the disturbance terms in any preceding year, we may use Aitken's
generalized least-squares and increase the efficiency ("efficient"
estimators have the smallest limiting variance and are necessarily
consistent and unbiased in the 1imit--although they need not be unbiased
2 Since the presence of serial
for finite samples?) of our estimates.
correlation3 and low intercorrelation of the independent variables
(lack of multicollinearity) is reasonable, we may usefully employ
Aitken's procedure. Fortunately, Zellner has developed a technique for
#Mood, op. cit., pp. 150-151; in other words, we may say ineffici-
ent predictors have needlessly large sampling variances.
2A. C. Aitken, "0n Least-Squares and Linear Combination of Obser-
vations," Proceedings of the Royal Society of Edinburgh, 55 (1934:1935),
pp. 42-48. It should be mentioned that two other consequences of auto-
correlated disturbances besides the inefficiency of the predictions in
the straightforward application of ordinary least-squares are that we
are likely to obtain a serious underestimation of the variances of the
regression coefficients; and although we shall obtain unbiased estimates
of alpha and beta, the sampling variances of these estimates may be
unduly large. See J. Johnston, Econometric Methods (New York: McGraw—
Hill Book Company, Inc., 1963), p. 179. '
3The well-known "regression fallacy" suggests that such things
as profit rates might be expected to exhibit serial correlation as the
extremes move toward the means.
81
using the Aitken generalized estimators when the disturbance variance
and covariance are unknown.1 This method allows us to make use of both
the time series and the cross-sectional aspects of our data. In testing
for significance, it is important to remember that the standard errors
are asymptotic standard errors. However, Zellner has shown that a
finite sample size of n = 20 (in our case n = 60), the results are not
very different.2 Our estimates using the "efficient estimators" ap»
proach is contained in Table 5.
We shall first look at the results on equation (1.1). Our results
are useful in answering a number of interesting questions. First of all,
the only variable that is significant is advertising expenditures. And
even here, it is only so four of the five years and even then it is ques-
tionable whether the level of significance is "reasonable" in all cases°
This suggests that our labors in computing profit rates on a fully ad—
justed basis was worthwhile. Advertising, and the constant term, have
a positive relationship to profit rates. Royalties and intangible as»
sets had negative signs for all five years. The sign for royalties
may be interpreted as a competitive profit-equalizing adjustment by
the capital market: the industries with small royalties have larger
1Arnold Zellner, "An Efficient Method of Estimating Seemingly
unrelated Regressions and Tests for Aggregation Bias," Journal of the
American Statistical Association, Vol. 57 (June, 1962), pp. 348-368.
2Arnold Zellner, "Estimators for Seemingly Unrelated Regression
Equations: Some Exact Finite Sample Results," Journal of the American
Statistical Association, Vol. 58 (December, 1963), pp. 977-992. It is
important to note that since the estimating technique does not mini~
mize the squared deviations around each individual regression, the
coefficient of multiple determination (R ) is not a relevant statistic.
82
TABLE 5
RESULTS OF GENERALIZED LEAST-SQUARES
Equation 1.1
~ Year N a1 b1 b21 b31
1956-57 60 6.893 8-024 -2.251 E-08° -1-243 E-07° 2.998 E-081
(5.379 E-03) (2.941 E-08) (1.459 E-O7) (2.215 E-08
1957-58 60 5.953 s-024 -2.202 E-080 -1.139 3-070 3.499 E-O8
(5.989 5-03) (2.979 E-08) (1.384 E-07) (2.093 E-O8)
1958-59 60 6.431 5-024 -1.780 E-080 -1.553 E-07° 1.126 E-08°
(1.187 E-OZ) (6.873 E-08) (4.129 2-07), (5.209 E-08)
1959-60 60 5.872 s-024 -1.764 E-080 -8.087 E-08° 4.500 8-084
(4.598 s-03) (2.252 E-08) (1.063 E-07) (1.673 8-08
1960-61 60 4.818 8-024 -4.224 E-O90 -7.481 E-08° 3.670 E-08
(5.411 E-03) (2.637 E-08) (1.053 E-07) (1.859 E-08)
Mean. 5.993 E-OZ 1.684 E-08 1.078 E-07 3.158 E-08
k
Equation 1.2
Year N 82 ‘ bzz b32
1956-57 60 6.828 8-024 -1.553 52-07o 2.886 E-081
, (1.303 2+01 (1.391 E-07) ' (2.184 E-08
1957-58. 60 5.888 E-02 -1.482 E-07° 3.457 E—08
(5.428 3-03) (1.329 E607) (2.062 E-O8)
1958-59 60 6.364 8-024 -l.826 13070 1.191 E-080
. (1 166 2-02) (3.818-E-07) (5.160 E-08)
1959-60 60 5.828 8-024 ~1.216 E-07° 4.618 8-084
(4.523 2-03) (1.017 E-O7) (1.660 E-08)
1960-61 60 4.812 8-024 -9.639 E-08° 3.806 E~083
(5.289 E-03) (1.024 E-O7) (1.835 E~08)
Mean - 5.944 E-OZ -1.408 E-07 3.191 E-08
Equation 1.3
*
Year N 33 b23 b33
1956-57 60 4.572 s+043 2.957 E-01° 2.361 E-02°
(1.792 s+04 (2.746 E-Ol (4.576 2-02)
1957-58 60 4.721 8+04 3.919 E-Ol 7.471 E-030
(1.834 s+o4 (2.259 E-Ol (4.085 E-OZ)
1958-59 60 .4.770 s+o4 3.580 E-Ol 5.927 15030
(1.806 E+04 (2.705 E-Ol (4.222 E-OZ)
.1959-60 60 . 4.513 s+04 4.362 E-Ol -1.336 E-02°
(2.040 s+04 (2.328 3-01 (4.591 E-OZ)
1960-61 60, 4.564 s+04 4.001 E-Ol -9.319 E-030
(2.086 E+04) (2.114 E-Ol) (4.579 s-oz)
Mean 4.626 E+O4 3.764 5-01 2.866 E-03
83
TABLE 5--Continued
*Standard errors appear in parentheses below the coefficients.
The E's are to be interpreted as indicating where the decimal should
be, +'s mean the present decimal should be moved to the right, -'s
to the left, by the number of places indicated by the number immedi«
ately following the sign, e.g., 6.893 E-OZ should be interpreted as
.06893.
not significantly different from zero
significantly different from zero at better than the 20% level
significantly different from zero at better than the 10% level
significantly different from zero at better than the 5% level
significantly different from zero at better than the 1% level
kth-IO
II II II II II
All tests are two tailed tgtests.
84
royalty-less profit rates to compensate. The results on equations (1.2)
and (1.3) are also interesting. They first indicate that very little
is lost by running the regression on profit rates without intangible
assets. Royalties is still never significant and advertising expendi-
tures is still significant in four of the five years with the signifi~
cance level for one year going from .10 to .05. Equation (1.2) further
reaffirms our suspicion that royalties and intangible assets are
positively related, while intangible assets and advertising are never
significant and with minus signs for two of the five years.
{/(After having found our profit rates, the next step is to find the
amount by which profits diverge from the "average." We may then add up‘7
all the pluses and all the minuses to find the amount of resource trans~
fer that would be necessary from low-profit to high-profit industries
to obtain equilibrium.1 Harberger estimated that profits from monopoly
power in the economy as a whole, which for him was manufacturing, summed
to $4.6 billion or 1.5 per cent of the national income (all in 1953
present value terms). In other words, the misallocation of resources
which existed in United States manufacturing in 1924-1928 might have
been eliminated by a net transfer of roughly 4 per cent of the resources
in the manufacturing industry or 1% per cent of the total resources of
the economy.
The question now becomes ”How do our estimates compare with
Harberger's?" Our answer is not one single answer but rather a combina-
tion of answers depending upon the method used. These are all shown
g
glAssuming that the elasticity is unity.
85
in Table 6. From our point-of—view, the most useful estimates are the
...... Tr~xNN\ g
ones involving after-corporation tax income with either of the capital
ases. If one wants to compare the figures with Harberger's, he used
ffi.
before-tax income on a total capital base. However, given the low tax
rates in the 1924-1928 period, it would be more instructive to compare
his figure with our after-tax income results. Before examining the
table, it is again worth mentioning that we have shown the misalloca-
tions based upon adjusted (for intangibles, royalties, and advertising)
as well as upon unadjusted profit rates. The Harberger figure mentioned
above refers to unadjusted data. His total figure after the intangible
adjustment, e.g., was 1.75 per cent of nationalincome.
The table indicates our spectrum or continuum of estimates ranging
from roughly $15 to $31 billion or from roughly 3.9 per cent to 8 per
cent of the average national income over these years.1 We can already
see that our figures are becoming of a different order of magnitude
than Harberger's (or Schwartzman's) estimates. These differentials will
become even more apparent after we apply the Hotelling formula to find
out how much better off people would be if we actually effected these
desired resource transfers.
n2
The summary results of our application of the "welfare formula
are shown in Table 7.
k
1The average for the years 1956 to 1961 is approximately $387.7
billion. Remember these estimates assume only unity elasticity. Our
results would be even more striking if we used the estimated elasticities
times "excess" profits measure of the required resource transfer.
2It is important to recall that we are using the wdrd "welfare"
loosely here to denote economic efficiency.
86
TABLE 6 ‘3//
ESTIMATES OF THE MISALLOCATION OF RESOURCES*
Profit Rate Methods
Using After-Corporation Tax Income of
Corporations (C) With Untaxed Partnership (P)
and Sole Proprietorship (SP) Income
(thousand dollars)
(U)
(I)
(1)9(R)
(1)5(R),(A)
Using Average Total
Capital Base:
(1) Estimated for
(P),(SP) by P
Balance Sheet
Data
(2) Estimated for
(P),(SP) by
Small Corpora-
tion Balance
Sheet Date
a. using $0-
25,000 total
asset class
be 1181118 $0-
50,000 total
asset class
Using Average Equity
Base:
(1) Estimated for
(P) 5 (SP) by
Balance Sheet
Data
18,931,308
(4.88)
20,354,724
(5.25)
18,765,574
(4.84)
14,989,845
(3.87)
18,999,768
(4.90)
20,453,334
(5.28)
18,862,497
(4.86)
15,080,923
(3.89)
19,554,572
(5.04)
20,995,162
(5.42)
19,414,998
(5.01)
15,630,044
(4.03)
24,674,585
(6.36)
26,441,728
(6.82)
24,630,725
(6.35)
20,947,663
(5.40)
NOTE:
equal, we shall always give the absolute average.
SOURCE:
See Table 2.
Since the sum of all the pluses and minuses are never exactly
Using Before-Corporation Tax Income of
Corporations (C) With Untaxed Partnership (P)
and Sole Proprietorship (SP) Income
(thousand dollars)
(U) (I) (UAR) \I),(R),(A)
24,249,442 24,360,424 24,765,897 30,828,524
(6.25) (6.28) (6.39) (7.95)
22,655,765 22,848,020 23,129,067 28,650,498
(5.84) (5.89) (5.96) (7.39)
21,104,421 21,356,222 21,873,126 27,329,557
(5.44) (5.51) (5.64) (7.05)
19,174,880 19,414,475 19,686,700 25,297,199
(4.94) (5.01) (5.08) (6.52)
*The per cent of average national income over the 1956 to 1961
period which the estimate comprises appears in parenthesis below the
estimates. The extent of the misallocation is defined here to be
the sum of the profits due to monopoly power in American industry in
the period 1956-1957 to 1960-1961 computed for unadjusted (U) and
adjusted for intangibles (I), royalties (R), and advertising expendi-
tures (A) data.
/
By assuming
unity elasticity is assumed.
4’
e misallocation equals profits means I)
-3.
5|
88
Although the general format of this table is similar to Table 6,
we should like to spend some time explaining the estimates, for in
essence, this is what our whole analysis has been pointing toward.
Table 7 looks at after~ and before-corporation tax income of C with
intaxed P and SP income. The estimates are done using both average
total capital and average equity as the base upon which profit rates
have been computed. The other modifications in the table refer to
the various ways in which the non-corporate sector's assets were
estimated.
The next thing to be noticed is that our estimates moved from
unadjusted to more and more realistic estimates.1 We make successive
adjustment for intangibles, royalties and advertising on the assumption
that these accounting items hide much of what is economically relevant
to the malallocation problem. For instance, we want to adjust for
advertising expenditures. For aldiough product prices might not be too
different from costs, the whole cost curve might be too high from
wasteful monopoly practices such as competitive (i.e., non~informative)
advertising.
It should also be noted that the estimates have been computed
using an elasticity of unity and using industry-by-industry elasticity
. . . . 2
estimates based upon the Lerner-Robinson prop031t10n. The reader
What per cent our estimates are of average national income appear
in the parentheses below the absolute estimates.
2The estimates based upon the Dorman-Steiner-Telser proposition
turned out to be such high ppper bounds to be worthless for our purposes.
Indeed, in some cases they indicated that the losses exceed national
income. Therefore we can move on to a higher indifference curve, or
TABLE 7
ESTIMATES OF AGGREGATIVE "WELFARE LOSSES"*
Computed for Unadjusted (U) and Adjusted for Intangibles (I), Royalties (R),
and Advertising (A) Profit Rates Assuming Unity Elasticity (451) and Lerner-
Robinson Elasticity Estimates (QL)
(thousand dollars)
Profit Rate Methods No (U) (I) (I) .(R) (I).(R).(A)
7? 1 71L 72:1 77L 74:1 77L 77:1 77L
Using After-Corporation Tax Income of Corporations (C) With Untaxed Partnership (P) and Sole Proprietorship (SP) Income
|'
Using Average Total Capital Base:
(1) Estimated for (P),(SP) by
P Balance Sheet Data 1 6,088,064 17,931,308 6,106,883 17,999,768 6,399,913 18,554,571 7,236,797 23,674,258
(1.57) (4.62) (1.58) (4.64) (1.65) (4.78) (1.87) (6.11)
(2) Estimating for (P),(SP) by
Small Corporation Balance
Sheet Data
3. using$0-25,000 total
asset class II 5,541,386 20,354,723 5,627,007 20,453,334 5,951,335 20,995,162 6,776,113 26,441,727
(1.43) (5.25) (1.45) (5.28) (1.54) (5.42) (1.75) (6.82)
b. using$O-50,000 total
asset class III 5,453,365 18,765,575 5,483,233 18,862,497 5,793,177 19,414,998 6,608,370 24,630,725
(1.41) (4.84) (1.41) (4.86) (1.49) (5.01) (1.70) (6.35)
Using Average Equity Base:
(1) Estimated for (P),(SP) by
P Balance Sheet Data IV 4,000,496 14,989,985 4,056,966 15,080,923 4,315,317 15,630,044 4,961,630 20,947,663
(1.03) (3.87) (1.05) (3.89) (1.11) (4.03) (1.28) (5.40)
Using Before-Corporation Tax Income of Corporations (C) With Untaxed Partnership (P) and Sole Proprietorship (SP)Income
f
Using Average Total Capital Base:
(1) Estimated for (P),(SP) by '
P Balance Sheet Data V 15,724,967 24,249,442 15,756,727 24,360,425 16,292,773 24,765,897 17,643,204 30,828,523
(4.06) (6.25) (4.06) (6.28) (4.28) (6.39) (4.55) (7.95)
(2) Estimated for (P),(SP) by
Small Corporation Balance
Sheet Data
a. using$0-25,000 total
asset class V1 10,398,164 22,655,766 10,442,740 22,848,020 10,816,925 23,129,067 11,855,472 28,650,314
(2.68) (5.84) (2.69) (5.89) (2.79) (5.96) (3.06) (7.39)
b. using$0-50,000 total
asset class VII 9,915,876 21,104,421 10,050,507 21,356,223 10,531,390 21,873,126 11,512,863 27,329,557
Using Average Equity Base (2.56) (5.44) (2.59) (5.51) (2.72) (5.64) (2.97) (7.05)
(1) Estimated for (P),(SP) by
P Balance Sheet Data VIII 8,919,432 19,174,880 9,000,850 19,414,480 9,346,280 19,686,700 10,299,487 25,297,199
(2.30 (4.94) (2.32) (5.01) (2.41) (5.08) (2.66) (6.52)
7
”Per cent of average national income over 1956 to 1961 period which the estimate comprises appear in parentheses below estimate.
SOURCE: See Table 2.
90
should also keep in mind Schwartzman's suggestion that perhaps an
elasticity of two is appropriate. It is interesting to note that
in at least one case, method V--we have numbered the estimates to
avoid repeating all the relevant information each time-wan elastiu
city of two would give larger losses than our estimated elasticities!
In order to give the reader an idea how these different esti-
mates are related, we have included the product moment correlations
(simple and partial) and the rank correlations. We have done this
for the fully adjusted estimates-~which are from our standpoint the
more realistic-~for both the unity (711) and (721) elasticity assump-
tions.
Regarding the correlations in Table 8, we find, for the most
part, results which our previous figures would have us anticipate.
However, there are a few interesting things to be noted. For
instance, although the simple product moment correlations under the
unity elasticity assumption are high, it is interesting to note that
some of the rank correlations are lower than the corresponding product
moment correlations. Also interesting is that, although the gwn
simple correlations of the absolute "welfare losses” and ”welfare
losses" as a per cent of business receipts are also high under the Ler-
ner-Robinson elasticity assumption, the simple correlations between the
higher income level, by specializing in leisure! This, of course,
does not affect their usefulness for the relative purposes mentioned
above. However, both the product moment and rank correlation co-
efficients for the various elasticities shown in Appendix C indicate
that the Dorfman-Steiner-Telser estimates may be limited even for
ranking purposes.
91
.Aoo n zv u wanna ca mocwauoo one moozuoa uNwoua economwfiv moons
uwwoum
uwwoum
uwmoum
uwwoum
ufimoum
ufiwoum
uwmoum
uwwoum
moan:
magma
wows:
wows:
wows:
wean:
wows:
wean:
NavaNme
..vaxNNvi
vaxNNevx
NaVNNNme
NNVNNNevx
NmVNNvax
vaxNAvi
. vaxNNvi
NeaN-oemN NmaN-emmN--mucmeamNNnmumm mwmeNmsm NN< sou muaNuuue mamaNmsm
NNN> eons»:
NN> venom:
N> vogue:
> 86508:
>N vogue:
NNN vogue:
NN oozes:
N 86:06:
mommog
mommoq
mommoq
mommoq
mommoA
mommoa
momwoA
mommoa
owmuo><
oummaos
summaoz
oumwaoz
oumwaoz
oumwaoz
enemas:
oumwaoa
oumwauz
"ouoz
ANNVx
Aevi
Nvax
Aevi
Avax
NNNVN
aNNVx
Aovi
Amvx
vax
Nvi
Nova
vax
Nevx
Anew
Nvi
fiNVx
w odnma
ca mcoNuoom Nam mam moanewue>
I .(III. II
Hz
vax
vax
Nvi
Noex
Amvx
Nevx
vax
Nvi
Nvi
«NauNua>
xwuumz ucowofiwwoou coNumNouuoo ucoso: uosmoum nomuo ouoN
coguaaomm< muwowumeNm huNcD “moumsnm4 haaam :Nmommoq oumuaozz
NewscNNaouv--m mamas
93
oooooo.H
mwmaem. oooooo.H
enamww. Macmew. oooooo.~
Naeamm. onaaam. whooom.
uwmmmm. moanwm. mmmwaw.
mmmaam. mmowmm. «mamam.
efiawam. wmqmmn. oowmsw.
mNNONw. amnesn. mwnmmw.
vax NNVx Nave
oonz
oooooo.H
mewmm. oooooo.H
mmmmmm. Hommmw. oooooo
Hmmcnm. mummam. omooom
mmmmmm. emanam. unmowm
vax Nevx vax
.H
. oooooo.~
. mmmaoo.
anx
oooooo.N
NNVx
vax
nNVx
Novx
vax
nevx
Amen
anx
Nvi
oNnaNNa>
:nommoa shamans: ousaomb< NOw xwuumz usowowmwooo cofiumaouuoo xomm
No.Nem u MM>
NN.oem m mN>
NN.aNm u N»
Ne.oom u 4 >
Nm.mmn u NN>
Ne.mNm u >
NN.omm nwfi>
NN.eae u >
NN.NoN u as
NN.NNN u w>
Nm.mNm u >
. ..Imw. I
. GHQ :13 l
Avuaauuaowvnnw mdm<fi
Nw.omm
NN.HHM
N~.mom
Nm.m-
N~.nom
N¢.oom
94
oooooo.H
wNwNmm.
Nmoenm.
mmmmmm.
Hmmmoe.
ooomom.
acmmmm.
manamm.
Nonmm.
chx
oooooo.~
Noommm.
emmmmm.
mmowen.
wwonwm.
acmwna.
memamm.
mmwmom.
omaaoa.
NmNQNO.
mwmaao.
Rococo.
oomwao.
m~m¢~o.:
omomoo.u
emmaoo.
smooHo.a
NNNVx
oooooo.H
mmmaeo.
mmonm.
ammomw.
Namomm.
mmNMNw.
mnmemw.
Hoeomm.
vax
oooooo.~
«mmmmm. oooooo.~
mmeoom. oflmmmn. oooooo.~
Hemomm. nmmmmm. mmomen. oooooo.~
owmmwm. ammoma. mmmmon. omwmmm. oooooo.~
Mnoemm. Nomoam. «NOHNN. omnwmm. wmaomm. oooooo.H
ONNon. cemmmm. ONNNMN. mwmmem. Cameom. mmnmem. oooooo.a
wmmmma.n HNmNmH.1 NHONHN.1 ceamma.n mowmma.u Onmmwa.u wmmmmN.n
unnono. Nomano. cammoo. monNH. mmmmaa. mmmqmo. momwea.
«ammeo. nooeeo. oo¢N~O.1 mmoHoH. enumeo. oommmo. omwooo.
mommmo. mmooeo. Hmamuo.u owoemo. madame. monwo. memeo.
Nowomo. genome. mwueofi. Haawoo. Nmomxo. wemmmo. Hameaa.
dwewao. mmmeao. omoN¢O.1 amamno. mammoo. mammoo.: Nemmmo.
Ammomo. someao. Deemmo.u Nameno. nooflno. HHmmqo. Homeaa.
mmmHNo. mwmoao. somomo.n mammoo. comamo. mmmmeo. mmnmoH.
ceumoo. Noomoo.1 Nmmeeo.1 NHNmoo. Romano. mmmmao. mqwaaa.
Novi Nvax Nevi NmNVx NNNVN NNNVN Novi
oooooo.~
mamnmm. oooooo.a
mmaemm. moomfim. oooooo.H
nammmm. unedma. ammnmq. oooooo.H
ommwwn. mowmwu. «Hosea. emmmoo. oooooo.H
momowo. «mmomo. Cenomm. awesom. mummwm. oooooo.H
«woaom. ammomn. endonm. wmoeam. HeHmmm. Newmmm. oooooo.a
Avi Nova Anvx Neva vax Nvi Nvi
NNNNN
Novi
Avax
Aevi
NmNVx
ANNVx
NNNNx
Novi
Aavx
vax
NNVx
Noex
Anvx
Nevx
vax
aNNx
ANVx
mNnaNsas
vax
“was
Nvi
novx
vax
Neva
vax
NNVx
NNVx
uNnaNnm>
xwuuex ucowowmmoou coNueNouuoo ucosox uosooum novuo ouoN
zuNoNuumHm conanQMuuocuuA wage: amouoanvd maflam :NmooooA unmwNoS:
NeoacNuaoov--m mam
:mommoq mummaozz ouSHOmn< you xwuumz ucowofimuoou cowumaouuoo xemx
NN.NNN nNN> Nm.NNN n 6>
Nm.Noe u NN> NN.6mm uNN> NN.emN u ms
NN.6me u 6N> NN.N8N noN> NN.o8N n 3>
Ne.aee u mNs Nm.6oN u m> NN.NeN u m>
Nm.Nmn u 3N> NN.8NN . as N3.3NN u N>
N6.Nmm u mN> NN.maN u N> Nw.NmN u N>
N
None N u: u > dogmas; mo mucowowmwooo any
NeonaNuaomv--w mamas
96
absolute losses and the relative losses are rather low: indeed, even
negative in some cases.1
To give the reader a more complete and detailed picture of what
we have tried to do, as well as providing some information perhaps
useful for other researchers for other purposes, we have shown in
Table 9 an industry-by-industry breakdown of our estimated losses
using profit rate Method I. This table shows how the losses for
each industry change as we make successive adjustments. In Appendix
D, we have included only the unadjusted and fully adjusted industry-
by-industry figures estimated by Methods 11, III, V-VIII, as well as
all intermediate adjustments under Method IV.
Returning now to Table 7, we can see that our “welfare loss"
estimates range from roughly $4 to $31 billion or from one per cent
to 8 per cent of national income. Compared with Harberger-Schwartzman
estimates of from one-thirteenth to one-tenth of one per cent of
national income, these are ratherlarge figures. The figure most
comparable to their estimates would be Method 1, fully adjusted, unity
elasticity.2
The estimate turns out to be about $7.2 billion or
1.9% per cent of national income (a non-trival change of magnitude in
itself). From our standpoint, we feel that any one of the after-
corporation tax, completely adjusted with industry-by-industry
1It would not have been worthwhile to show the partials of one
profit method on another as the high intercorrelations made the matrix
approach singularity, i.e., the ith variable to be added was found to
be approximately a linear combination of the previous i-l variable.
This same multicollinearity shall prevent us from showing partials on
different elasticities, concentration ratios, etc.
2Actually he made no adjustments for royalties.
97
TABLE 9
INDUSTRY-BY-INDUSTRY "WELFARE LOSSES" FOR PROFIT METHOD I INCLUDING ALL INTERMEDIATE ADJUSTMENTS*
Unadjusted (U) Profit Adjusted for Adjusted for (I) and Adjusted for (I), (R)
Rates Intangibles (I) Royalties (R) and Advertising (A)
75: 1 71-2. . 71: 1 “Yb n: 1 “Yb. "VP 1 71L
Industry
1. Agriculture, forestry and
fisheries 131,778 1,469,836 132,674 1,474,825 143,078 1,531,564 167,100 1,655,146
2. Metal mining 11,947 113,093 11,637 111,615 12,206 114,313 12,277 114,645
3. Bituminous coal and lignite
mining 4,171 70,732 4,156 70,610 4,348 72,217 4,470 73,229
4. Crude petroleum and natural gas 29,064 292,815 28,091 287,870 29,289 293,941 31,080 302,800
5. Mining and quarrying of non—
metallic minerals and
anthracite mining 10 3,137 25 5,056 41 6,482 125 11,362
6. Construction 24,626 784,619 24,749 786,582 25,489 798,260 33,469 914,720
7. Beverage industries 436 42,075 460 43,222 526 46,242 14,923 246,243
8. Food and kindred products 423 101,235 441 103,352 531 113,432 14,148 585,250
9. Tobacco manufacturers 65 12,125 65 12,204 90 14,283 5,595 112,841
10. Textile mill products 4,103 166,477 4,103 166,482 4,393 172,252 7,095 218,920
11. Apparel and other finished
products made from fabrics
and similar materials 36 14,190 38 15,264 50 17,529 969 77,128
12. Lumber and wood products
except furniture 500 46,259 680 53,972 780 57,770 1,112 68,981
13. Furniture and fixtures 3 2,944 5 3,555 9 4,808 573 38,095
14. Paper and allied products 563 55,987 574 56,523 704 62,951 2,115 108,474
15. Printing, publishing, and
allied industries 469 55,177 579 61,291 667 65,774 1,862 109,930
16. Chemicals and allied industries 252 55,517 299 60,541 442 73,597 23,219 533,420
17. Petroleum refining and related
industries 36,176 809,850 34,299 788,563 36,309 811,343 44,706 900,279
18. Rubber and miscellaneous
plastic products 180 24,794 182 24,916 220 27,418 2,041 83,470
._._., .., »;L__;a.agi;;;_’:£i’f§i“ififi'fifii '.;—.-‘73"'.*.rar.n_z.:.'$'.§.. -...... . . _. . . -... . .. .. - .. .... .I... ..-. I ,. ... . -.-. . , ,. __ -. , . . .........,____i ..
98
TABLE 9--(Continued)
Unadjusted (U) Profit
Adjusted for
Adjusted for (I) and
Adjusted for (I), (R)
Rates Intangibles (I) Royalties (R) and Advertising (A)
Industry
71: 1 74L 71‘: 1 72L 71: l 77L 74: 1 71L
19. Leather and leather products 172 18,060 173 18,096 193 19,106 870 40,582
20. Stone, clay, and glass products 34 12,962 38 13,622 70 18,426 624 55,148
21. Primary metal industries 4,167 237,238 4,186 237,769 4,720 252,485 6,942 306,207
22. Fabricated metal products
(including ordnance)
except machinery and trans-
portation equipment 935 93,238 952 94,101 1,099 101,061 3,720 185,962
‘23. Machinery, eXCept electrical
and transportation equipment 684 94,875 729 97,949 919 109,916 4,740 249,667
\24. Electrical machinery,equipment
and supplies 73 27,265 90 30,333 135 37,189 4,549 215,572
25. Transportation equipment
except motor vehicles 9,265 271,698 9,325 272,574 9,587 276,376 11,042 296,612
26. Motor vehicles and motor
vehicle equipment 929 105,335 977 107,982 1,143 116,834 3,977 217,904
27. Professional, scientific, and
controlling instruments;
photography and optical goods;
watches and clocks 20 7,594 27 8,817 44 11,167 2,099 77,265
28. Other manufacturing 748 53,787 756 54,071 828 56,579 2,865 105,255
29. Transportation 145,772 1,529,379 145,387 1,527,362 152,948 1,566,574 167,984 1,641,773
30. Communication 11,684 275,051 11,847 276,962 13,275 293,182 16,634 328,177
31. Electric and gas companies
and systems 146,795 1,097,966 146,839 1,098,134 156,602 1,134,049 162,015 1,153,483
32. Water supply and other
sanitary services 7,575 37,911 7,566 37,866 8,029 39,030 8,123 39,257
33. Groceries and related products 198 57,378 202 58,015 221 60,726 950 125,799
‘34. Electrical goods, hardware,
and plumbing and heating
equipment 73 22,852 75 23,202 94 26,049 573 64,244
TABLE 9--(Continued)
99
Unadjusted (U) Profit
Adjusted for
Adjusted for (I) and
Adjusted for (I), (R)
Rates Intangibles (I) Royalties (R) and Advertising (A)
Industry
74:1 WL 74:1 72L 72:1 77L 77:1 71L
35. Other wholesalers 861 208,066 891 211,572 1,056 230,454 6,196 558,266
36. Food 3,031 274,936 3,096 277,881 3,253 284,843 9,634 490,190
37. General merchandise 122 39,849 126 40,475 182 48,763 178 48,195
38. Apparel and accessories 172 31,826 179 32,445 217 35,707 5,320 176,926
39. Furnitdre, home furnishings,
and equipment 16 8,705 21 9,863 37 13,189 5,957 168,104
40. Automotive dealers and
gasoline service stations 19 21,787 22 23,727 43 32,915 2,693 261,624
41. Eating and drinking places 9,193 256,286 9,556 261,300 9,831 265,032 15,389 331,597
42. Building materials, hardware,
and farm equipment 235 38,748 238 38,968 294 43,302 1,368 93,400
43. Other retail stores 4,823 249,663 5,083 256,308 6,993 87,703 14,736 436,393
44. Wholesale and retail trade
not allocable 105 59,812 112 30,852 5,396 264,068 1,671 119,130
45. Banking 865,763 799,413 881,804 806,/83 147 35,281 1,105,018 903,143
46. Credit agencies other than
banks 3,257 59,852 3,446 61,571 927,337 827,353 20,993 151,958
47. Holding and other investment
companies 259,917 126,403 265,103 127,657 325,497 141,453 332,932 143,059
48. Security and commodity
brokers, dealers, exchanges:
and services 20,423 133,057 20,679 133,890 22,493 139,639 57,591 223,439
49. Insurance carriers 3,632 190,834 3,666 191,724 4,233 206,006 5,239 229,188
50. Insurance agents, brokers,
and service 109,974 433,321 110,062 435,459 111,682 436,674 123,762 459,682
51. Real estate except lessors of
real property other than
buildings 1,501,999 2,034,105 1,498,593 2,031,797 1,582,567 2,087,948 1,806,252 2,230,631
52. Lessors of real property,
except buildings 1,209,973 161,487 1,206,589 161,261 1,258,229 164,675 1,262,879 164,979
100
TABLE 9--(Continued)
Unadjusted (U) Profit
Adjusted for
Adjusted for (I) and
Adjusted for (I), (R)
Rates Intangibles (I) Royalties (R) and Advertising (A)
Industry
77:1 71L 7Z=1 71L 7751 72L 72:1 73L
53. Hotels, rooming houses,
camps, and other lodging
places 26,065 227,589 25,942 227,052 27,306 232,942 39,448 279,983
54. Personal services 47,470 422,295 48,036 424,805 48,887 428,550 65,161 494,764
55. Business services 17,057 285,934 17,409 288,870 17,800 292,095 23,160 333,182
56. Automotive repair services,
and garages, and other repair
services 20,091 249,964 20,239 250,883 20,729 253,905 28,104 295,640
57. Motion pictures 5,720 86,529 5,509 84,918 5,783 87,004 13,621 133,528
58. Amusement and recreation
services, except motion
pictures 150 14,487 291 20,178 328 21,418 2,420 58,150
59. Other services 1,402,974 3,456,046 1,405,838 3,459,572 1,409,326 3,463,860 1,520,728 3,598,160
60. Nature of business not
allocable 1,096 26,143 1,123 26,456 1,189 27,230 1,758 33,110
TOTAL "WELFARE LOSS" 6,088,064 17,931,308 6,106,883 17,999,768 6,399,913 18,554,571 7,236,797 23,674,258
rates for C are average returns over the five-year period 1956-
*This is computed for all business establishments—-Corporations (C), Partnerships (P). and Sole Proprietorships (SP)~-using after~C
tax income of C with untaxed P and SP income and estimating P, SP average total capital figures by utilizing P balance sheet--the profit
1961 for P, SP.
SOURCE:
IRS Statistics of Income, Corporation Income Tax Returns, and Business Income Tax Returns for the relevant years.
3957 to 1960—1961, while they are for four-year period 1957-1958 to 1960-
101
elasticities,estimates are the most relevant for estimating effici-
ency losses. These estimates run from $20.9 to $26.4 billion or
5.4-6.2 per cent of national income. We discussed in an earlier
chapter what we felt were the analytical and empirical drawbacks of
the unity assumption. To be accurate, we can trace our rather sharp
increase over the Harberger estimate to two main sources: the change
in scope and the change in elasticity assumption. Of course, there
were also a number of other, less-important,influences.
To get a better idea of the order of correlation between
Harberger's estimates for the 1924-1928 period and our estimates by
Method I for approximately the 1956-1967--1960-1961 period,as well
as to establish a number of interesting sidelights, we have included
our product moment correlation results in Table 10. We were able
to compare our findings by lumping the seventy-three manufacturing
subindustries Harberger used into our twenty-two industry SIC
schema. It should be noted that a number of our correlations are
spurious, in that one of the arguments is partially composed of one
of the others, e.g., X(lO) and X(3). However, to preserve the
symmetry of the correlation matrix we have presented the entire
results.
The important correlations for our comparison are R10,9 and
R11,12--the relationship of Harberger's "excess" profit rate esti-
mates, X(9), to our estimates, X(lO), and his estimated "welfare
losses" divided by sales, X(ll), to our ”welfare losses" divided
oooooo.u onx
«muses. oooooo.~ Amcx
«Naume. mo3m33.- oooooo.u Asvx
Nmuoam.- ommowu.- mammoH. oooooo.u Amvx
wmosua. ammusa. ~H~n43.- somaua.- oooooo.u x~cx
Hemne~.- meao¢¢.- ommONH. owuamm. mm~030.- oooooo.u Advx
onx Amvx asvx Anvx Amos Aavx oHanHm>
xwuum: ucmwofiwwooo coHumHouuoo ucoEoz uosvoum umpuo oumN
102
nonwoomm mmocwwsm mmmuo><\:wommoq mummao3: Mac n “Navx
moHom\:mmmmoA cummaozz m.umuuonumm u AHHvx
Qumm uamoum :mmmoxm: H30 " AOHVN
mama uwmoum :mmmuxm: m.ummumaumm n nmvx
Homu-ooa3 on nmmfluemma mcuusuuamscmz an mudumumm mascumsm mwmum>< u vax
Homanooma ou mmmauommHuuH vogue: uwmoum ha wowusuommscmz aw :mmmmOA mummamZ: wouoawumm uso u Auvx
wmmu you wantsuomwacmz an mmfimm u onx
mNmH-¢~aH masuzuumwsamz as zmommon atmoamz= m.umwtmnum: u Amos
Homa-oomu on anm3-ommu not wantsuuamsamz an umuuaao sauce mwauu>< saunasumm “so u Aevx
Homfluooaa cu ams~-ommu--H cocoa: osmoum an mastsuuaunaaz as mascots =mmuoxm: umuasuumm use u Amvx
mNmHuemmH .wdwuauowmssmz cw amuwamo Hduoe w.uowumnunm u Amvx
mamasemma «wcwuauUdmacm: cw muwwoum :mmooxms m.umwuonuum u AHvx
moanmwuu>
:mmmmOA ma Ne.m3m u m> No.003 u o> so.omm u m>
sm.N33 "33> s¢.~3a u m> so.w¢3 n m> sa.om3 n ~>
so.¢~n "03> sm.mm~ u as s¢.mNN n 4> s3.mme.m u 3>
_x
”mum .Hum. u > «cofiumwum> mo mudmfiuwmwooo 05H
.m.p now roundups wumuwpafi muons
em. u m..~m. n m
..
omuuAm.pv Eonoouw wo moouwmp aNN u z 00. u mm.«oH. n NM NH Ham
.HU>UH AUNONV Nm Gun: kuumn um OHON Scum uflouommwv mm. H M «We. ..u m
s m «3.33 m.o3 . -1 . u . . u a.o3
33aao3u3c 3a m3 A «V m momma 3 em33qu can wa3m= «3 ..3m. w3 we a
oooooo.3 A~3ux
-a3~m. oooooo.3 A33vx
moosm¢.- scemNM. oooooo.3 2033x
~¢w~3m.- mowaoo.- aaNANe. oooooo.3 Amvx
«m3oem. wsmNo3. o~w-m.- moa3¢~.- oooooo.3 Amos
333m4m. m3q3m3. w3mwmm.- encom~.- ammo~¢. oooooo.3 “sex
o303mN. ommsmo.- mmmmmm.- 33wmo3. «comma. amommn. 303x
a3moo3. qmmomm. ma3mm~.- ammoao.- «amass. m33aa3. Amvx
soma33.- aeomm3.- «amoe3. sweamm. mom03~.- mammwo.- Aevx
~m¢-o.- 3044mm. oam3mw. 3maomm. e3aamq.- assoaa.- Anus
m~3¢~o. «maqe3. oammmo.- m--~.- mamoem. «meemo. “Nos
a3oa¢¢.- anneao.- smoqae. oa333a. Aesmm~.- mm~3q¢.- x3vx
x~3vx x33vx Ao3vx Am3x chx Aavx ~33u3uu>
Avo=a3uaoovu-o3 mamas
104
by business receipts, X(lZ).1 Only the first of these two correla—
tions, R10,9, is different from zero at any reasonable level of
significance, 5%. 'nxzother is significant only at the 20% level--
both being for two-tailed 5 tests, where 5 refers to "Student's”
distribution. We divided the losses by sales (or business receipts)
to remove the scale factor, For, we did not want the mere growth in
size of an industry and hence the possible growth in absolute size of
the "welfare losses"--proportionate losses remaining the same-~to
indicate a growth in losses. Later, we shall discuss changes in
absolute losses, but for now we are concerned with relative losses
in our estimates vis-a-vis Harberger's. For what it is wbrth, using
absolute, rather than relative, "excess" profits ("welfare losses")
yields much higher (lower) correlations between his estimates and ours.
V
1Since we shall be making use of both R 2 (read R bar squared)
as well as R2(the coefficient of multiple determination), it is useful
to spend a moment distinguishing between these two. Since R2 is
defined as the sum of squares"explained by"(due to) the regression
divided by the total sum of squares, 100 - R2 is the percentage Of
the sum of squares of Y "explained by" (or associated with) the indepen-
dent variables. Since'R2 = 1-(l-R2) 523.3..(where N = number of obser-
N-k-
vations, k = number of independent variables; alternatively we may
think of the latter term as N'1 where m = number of degrees of freedom
N-m
used up in fitting the regression equation), it is smaller than R2 for
any finite sample size. (Incidentally, if desired, the coefficients of
partial correlation may be adjusted using the same formula). All this
means that R (R2) is biased upward for small samples while'R (or R2) is
unbiased. In terms of our previous phraseology, we may say that since
R2 gives the exact split of variance into explained and unexplained
variance whereas R2 splits the sum of squares, 100 -'R2 is the percen-
tage of the variance of Y "explained by" the independent variables. In
simple terms, we may say that since adding another independent variable
to an equation can only increase, or at the limit leave unchanged, R2,
a researcher could add variables ad infinitum until some "desired”--
presumably high--"goodness of fit" was obtained. Hence, we want to
105
Of course, for our more refined estimates, the correlations%‘
between our respective estimates would be much lower. Before moving
on there are a number of interesting relationships indicated in the
preceding matrix that are worth spending a.moment on;1 for instance,
the negative relationship between the "welfare loss" ratio and the
"excess" profit ratio (and absolute amounts) and with the average
total capital variable. Remembering that the minus profit rates are
computed by subtracting from an overall average, i.e., the under-
production in minus industries is relative to overproduction in "high"
profit industries, the relationship is not so surprising. In other
words, since Our individual "welfare" estimates show the amount by
which consumer "welfare" would increase if that industry either
acquired or divested itself of the appropriate amount of resources,
the negative sign merely indicates that the desired minus resource
reallocation is not exactly matched by the plus transfers in this
particular case-~by some other estimating methods we get the opposite
sign. The reversal of sign in one case when "welfare losses" are in
absolute terms for the second relationship is also suggestive that
the negative sign may be ambiguous. The partial correlation coeffici-
ents indicated that the positive correlation of "welfare losses" and
sales (business receipts) combined with the high negative correlation
k
attach a "cost" to the addition of variables to weigh against the
"benefit" of a higher per cent of "explanation? \Ne do this by de-
creasing the degreeg of freedom by one each time we add a variable.
Thus, an enlarged‘R is more meaningful than an enlarged R2 and
corrects the upward bias of R2 for small samples.
1The reader can gain some insight from the measure of relative
diapersion, the coefficients of variation or the estimated standard
deviations divided by the estimated means, which we have included at
the bottom of Table 10.
106
of sales and "excess" profits explains part of the sign as well.
The "partials" between ”welfare losses" and "excess" profits becomes
less negative as we add average total capital (ATK) and business re-
ceipts to the equation. It is also interesting to note that the "par-
tial" between ATK and"welfare losses" changes from negative to positive
after business receipts is added to the equation with "excess" profits.
Similarly, the "partial” between "welfare losses" and "excess" profits
falls slightly as ATK is added. Nonetheless, the implication that the
smaller the industry, in asset terms, the larger the "welfare losses"
(in absolute or ratio terms) is certainly interesting though hardly
conclusive. For one thing, since the large firms may be able to hide
"quasi-monopoly" elements better, our adjusted figures may be more
relevant. And, of course, such things as economies of scale are
relevant in this context.
In Table 11, we have shown the rank correlation coefficients
for the two estimates of absolute "welfare losses" and "excess" profits.
The correlation for ranks appears more significant than the
product moment results. The reader should be careful to note,
although R has approximately the same value as R in Table
3,4 11,12
10, we are now comparing absolute "welfare losses" and not "welfare
losses" as a percentage of sales--the former, before was only
R7,5’
.18 not .32. However, this suggests that the significance level for
the latter would also be higher. Incidentally, the negative correla-
tion between "excess" profits and "welfare losses" for the zero order
coefficients disappears for one of the "partials." Thus, R1’3.24
[variable 1 on 3 with the effects of (dependence on) 2,4 "taken out”]
107
.< aquamaa< van u wanna aw conwuumov H wonuoz uwuoum hm "mumaom
.~o>oH A.m«w NH eonu Houuon um onus aouw uco
. . a n
uuowmwm haucmofiwwcwwm ma A My N am mumounm voawmu osu mean: u m cm NN z
mm. a m. an. n a
a
so u mm. as n mm a mm
mm. u _m as. u a
*Nl Nag
.u.e How eouasfiea massaeaa enema em. mamm. an. n Nu m
oooooo.s Asvx
Namssm. oooooo.~ Amvx
ooo~n¢.- wmmemo. oooooo.H Amvx
amaomm.- mamaeo. «comes. oooooo.s AHvx
Aavx Anvx Amvx Aavx usnmaua>
Xauumzvucowowwmooo sawumaouuoo ucofioz uosvoum nacho ouoN
Asux
Homauooma cu mmmH10mmH you mucoazmwaaouwm wcwunuomwscmz you :mommoA mummaoS: uao u
wumau¢~ma you acowuouoauoo wcwunuoowacmz cw :wommoq oummaoza m.uowuonumm u Amvx
Homauooafi ou unmanomma you mucosnmwanmumu wowusuomwocmz How muwwoum :mmouxm: use u Auvx
wmmansmma How seawumuonuoo wcwusuuomscm: cw muwmoum :mwooxm: a.mowuonus: u Aavx
moanowuo>
:mmmmoa mmdfidmza 92¢ mHHthm ammuoxw: QHH<
acoocfipomnuoouoa an zuwoaummao on wowanmum :amoumoq.oumwao3: voumofivm madam
«hugs: mo muwoauaoao on mowanamo :amommou ounufioaz voumonum haaom
acomownomauoouog an huuouummao on wdwanmmm :«mommoq ouomaoB: womanhood:
thugs: mo huwowumoao on wowesmmo :amommoq mummaozz moumsmvmca
taunt comments madam mean: can hugs: mo huwoaumuao on wowanmmm peace maonoooe mo novcH a.moouoa
utmost voumonvm mason wagon vow hows: mo huwoauumao so woweommm meson haoaooofi mo xovcH a.mosuoa
trauma voumsnvoon wage: was hugs: mo muwoflummao on mafiaommo uoaom maomooofi mo xovoH a.moauoq
«oaauooma ou unmauomma wow oumu uwuoum
Homaaooma ou mmmfinomma now mum» ufiwoum
oHumu :oHumuucoocoo ucoahoAmBo
owumu coHumuuaooooo ucoahoamfio
ofiuwu cofiumuucoocoo ucm5m04oam
oHumu aofiumuusoucou uooshoamao
oHumu co«uwuucoocoo cotteue:.n:
owumu newumuuoouaoo movmnnooam>
owumu co«uouuooocoo coconuo=~m>
oHuou cOwunwuoooooo covvmaoo~o>
unwavuw
uamHv-~
newne-~
ssmau-~
uawfiv-~
unwac-m
uamHv-~
unwav-~
AwNVx
As~vx
Aouvx
Amavx
Annex
“havx
Aoavx
Andes
Aedvx
Ansvxx,
Mmfixx
as x
AoHvth
flax/r
Amvx
Aavx
Amen
Amvx
“sex
Amvx
Aavx
Amen
L '0“. l . Ill . 14.| l‘llllllll1|l’fl.l| . [1|"1ll'llu ’I'Illull 1-.-! ' '.| I- 041’ ll! .
moannwun>
:mmmmoa NMdNAHS: nz<_mm30m MAOAOZO:
as xmnza a.muzama .mmaae aHmoam .mouaea onaaeazuozoo aHuHa-oza so monaagmmaoo azuza:_au=ooem
NH mamas
oooooo.H a-vx
macaam. oooooo.s AHNVx
owaeao. aoNamm. oooooo.H Aouvx
mmoamq. ammmmm. Nausea. oooooo.~ Amsvx
ennmso. seamen. wamwes. asae~m. oooooo.~ Annex
~ms~me. «Heomm. mammom. «amuse. wosmmm. oooooo.s Aaavx
~amsoe. aoemsa. smegma. cosmos. «scams. «amass. oooooo.a Assvx
mnmmmm. meommm. mmomma. HNNosm. seesaw. namema. oaeamm. oooooo.s Annex
a~s~mm. seamen. osuwma. msmnem. namn~s. manna“. “comma. summam. oooooo.H Asses
enemas. mesons. amsomo.- mamman.- s~omm~.- m~ee-.- mwamao.- omomm~.- anaem~.- oooooo.s Annex
ommsmN. manoMN. ommaam...omomoe.- smooo~.- emNmm~.- Hm~aa~.- wamsn.- mmmmm¢.- tomes». oooooo.a a~avx
swmsmo. -m~mo. ammomn.- mamanm.- ~mam¢~.- mooaem.- msomam.- moosma.- ma¢am¢.- “macaw. monomm. Annex
momasu. masomN. ameMs.- HnNNmH.- ~oamao.- masmeo.- memsas.- ~0H-m.- awamsm.- NNNaNm. HeaQNm. Aosvx
Hmsoaa. meomHH. ounces. hennoo. ma~mn~.- ~m¢oss.- aomm~a.- «moons.- Hmmmss.- “sham“. smomON. Asvx
~o~oso. sassaa. ommmes. onNHe. «scams. magmas. masons. ~o~mam. smmamm. ssmamm. «macaw. Amos
meshes. macaNN. manage. Hansen. nmsmms. osmmae. mosses. asmmam. ossmsm. cammmm. season. Aavx
mesons. mmmoaa. meoNem. mooman. nausea. seamen. seamen. mmmeQN. mmame. mammae. Namoas. Aces
smammm. mmmmuo. aweumm. musesw. ammoao. summom. Nomoam. «amass. NHNsos. macame. mwmsos. “max
mmwseo. mammoa. saan~s. mosses. ammosN. unease. «scams. Hamaos. Newman. Namamm. mNNmNN. Aevx
sesame. anomNN. mamaam. seawam. NnmoNN. accuse. mssmms. unwaom. ammmmm. «mason. maowsm. Anvx
"u canons. enemas. stooNN. Haansm. ooNaoH. ammscm. mmmwam. macasm. msmwms. omamas. ewmmma. “max
.1 mmoomm. o-aoo. masons. Nanama. ommamo. Hmooou. -mmms. mmmamo. assume. mosomm. NNmsnm. Anew
x~mvx AsNVx Aowvx Asses Awsvx Annex Aoavx Amsvx Asses Annex Amavx assessm>
oooooo.H Annex
osomsm. oooooo.H Aosex
omsooa. emsmew. oooooo.s Assn
samNNo. ascoem. manoNN. oooooo.s Amvx
mHonH. stwom. oaaaw~. canoes. oooooo.s “aux
Naowmm. nausea. oamamm. -o~om. ssmamm. oooooo.H “sex
«mawmn. Nooosa. saom~e. -aemm. HHNsom. Humans. oooooo.~ “max
masses. HmemN. mammas. mwe~mm. Manama. demons. mso-¢. oooooo.s Aevx
omosma. mmmomm. seahoN. mmmmmm. oaanam. chases. moNNsm. ~Hssmm. oooooo.s “aux
mNNomN. canons. wasasm. m¢~ssm. sosmom. magmas. nausea. scammm. enemas. oooooo.H “Nos
Nsmomm. mNsmom. osness. emmamm. semmmm. omsmam. masons. enemas. access. masses. oooooo.s have
Assvx Aosvx Amvx Amos Aavx Asvx Amvx Assn Amvx Amos Aavx ossaauo>
“Shun: oofioaouuou uooao: nonvoum macho 9:5
1 {I'll-'1'"- ‘llll'llllo .IIAII’ I 'I‘ --l'll. ‘1".I'l.'l' [“1 l,|.l‘1.|.f.|.ll.,|1’|1 .ul 1'" .ll‘l
Aconcaucoov-~a mamas
112
too few resources. This means that an industry could have large losses
and be highly competitive since its below average profit rate minus the
overall profit rate,~the per cent of underpricing, yields a negative
figure. But when squared for determining ”welfare losses" it may yield
a higher "loss" than a less competitive firm--albeit for different
reasons. In short, big "welfare losses" are not necessarily indicative
under our model, that a firm is highly monopolistic. What is more rele-
vant for comparison with evil effects of monopoly positions as reflected
inggR is the percentage that prices are too "high" or too filowf com-
pared with an optimalrallgcation of resources, i.e., the Lerner index
w
under our assumptions. This latter index retains the sign of the per-
centage deviations so that we may usefully compare these with CR. Since
some readers will undoubtedly be interested in the ranking of industries
by this criterion, we have included our industry rankings, from highest :
to lowest, computed from profit Methods I and IV for FA,‘Tz= 1 data in
Appendix F.
Returning to Table 12, we find that the Lerner index, without
minus adjustments,1 based on FA profit rates, 7? = 1, has, without
exception, a higher relation to CR than does U profit rates. This
again reaffirms our belief that the adjusted results are much more
realistic for determining the malallocative effects of monopoly posi-
tions. We also ran the Lerner index for the FA results,'7l= l, with
minus adjustment, i.e., the method we employed in computing "welfare
1For a detailed description of the distinction between with and
without minus adjustments, see Appendix A's section entitled "Profit
Rate and ‘Welfare Loss' Adjustments."
113
losses." As we might expect, the results are slightly less clear-cut--
the with exceeds the without correlation with CR in all but one case.
Our results for absolute "welfare losses" and relative "welfare
losses" follows much the same pattern as the above. That is, the FA
results are more significantly related to CR than the U. The reader
may spot check this rather significant difference by comparing the U
results of X(lé), X(lS) and X(19), X(ZO) against the corresponding FA
results of X(16), X(l7) and X(21), X(22). The only result which is
'hegative" as far as our analysis is concerned is when our 71 L' esti-
mates are less significant than the'2> = 1 results on CR. Except for
X(l4), X(lS) and X(l9), X(ZO), the coefficients are reversed from what
we might expect. However, since the magnitudes are in the proper
direction in the other cases mentioned above, there is the possibility
that randomness could account for our unexpected results.
Turning the emphasis around, the reader can take a given estimate
assuming it reflects the "true" relationship and examine the simple
correlations of X(l) to X(8) to see which of the CR beSt estimates
the "true" relationship. There does not appear that any one Of th? CR
is "best" under all conditions. In other words, no one of the CR, say
value-added of the four largest firms, always yields the largest R for
the different estimates.
CHAPTER V
SUMMARY AND CONCLUSIONS
This dissertation has consisted mainly of a further development
of the Harberger-Hotelling technique for the estimation of "welfare
(or more accurately efficiency) losses." The further development has
been concerned with extending the scope and timing of the empirical
investigation and modification of the theory to permit a more realistic
appraisal of the general order of magnitude.
Our investigation of the "welfare losses" in the American econo-
my extended over approximately the 1956-57 to 1960-61 period. We
first obtained estimates of profit rates for corporations (C), partner-
ships (P) and sole proprietorships (SP), per annum, over the five-
year period 1956-57 to 1960-61 for C, and the four-year period 1957-58
to 1960-61 for P and SP as well as the average rates over this time.
This was done on the basis of unadjusted data and data adjusted for
royalties, intangible assets and advertising expenditures. Further-
more, the rates were computed using both before-tax and after-tax
income and using both total capital and equity bases. We also com-
bined our above results to obtain the average profit rates for all
business establishments. This was the information that was eventually
used in our first approximation to the "welfare losses." The rank
correlations between our eight basic profit methods were quite high--
Ro’u‘264 being the lowest coefficient.
114
115
We discovered, by the "efficient estimators" technique
(generalized least-SquareS) that advertising expenditures was
the only variable among the reputed monopoly indices of royal-
ties, intangible assets and advertising expenditures that was
significantly related to the unadjusted profit rate. This
suggests that our computation of adjusted profit rates was worth—
while. Further regressions indicated that little is lost by
dropping intangibles out of the equation and while royalties
and intangibles are slightly related, intangibles and advertising
expenditures are never significantly related. Our a priori
suspicion that royalties and profit rates may be negatively
related because of competitive "returns -equalizing" forces in
the capital marketwas substantiated.
We further observed the amount of resources that it
would be necessary to transfer to equalize profit rates (assuming
unity elasticity) ranged from roughly 3.9 to 8.0 per cent of
national income as compared to Harberger's estimated 1 1/2 -
1 3/4 per cent of national income. However, before we could
utilize this information for estimating "welfare losses" we
needed to know something about demand elasticities.
We estimated elasticities for all our industries by two
different methods. One method which employed advertising inten-
sity data yielded upper bound estimates which were of little use
for our purposes. However, the estimates which utilized the fact
that the reciprocal of Lerner's index of monopoly power yields \\
an estimate of actual (not upper bound) elasticity, provided we
116
have a profit-maximizing firm in equilibrium,were an important
addition to our analysis. We estimated these elasticities separately
for all our "welfare loss" estimates since changes in "excess"
profits affect these estimates. The product moment and more espec-
ially the rank correlation coefficients between 232 elasticities
were quite high, i.e., the relationship was significant as betWeen
different estimates of either the advertising or marginal estima-
tion method separately. However, the negative correlations between
the different approaches suggests that we must be careful about the
utilization of the advertising approach foreven relative ranking
purposes.
Utilizing our estimated elasticities as well as employing
the Harberger assumption of uni(:)elasticity and the Schwartzman
assumption of an elasticity of two, we calculated "welfare losses"
that range from one to eight per cent of national income. Previous
studies placed the losses in manufacturing around one-tenth to one-
thirteenth of one per cent of national income. Even using approxi-
mately the same methods and assumptions as Harberber, we get losses
in the vicinity of two per cent of national income for the whole
economy. This suggests that either the losses have increased in our
investigated time period relative to his period or that our estimation
procedure is more inclusive.
Further investigation of Harberger's "excess" profits and
relative "welfare losses" with the estimate which most nearly para-
lelled his, we found a significant relationship between only the
first of these variables. To impart some flavor to the reader as to
117
how the magnitudes change as the various adjustments are carried
out, we have shown some detailed, step-by-step, estimations by the
two most realistic profit methods. We further noted that although
most of the "welfare losses" are highly interrelated, the absolute
"welfare losses" and the relative "welfare losses" (i.e., "welfare
losses" divided by business receipts) had a low correlation--indeed,
negative in some cases. The rank correlations on the absolute "wel-
fare losses" were somewhat higher. An observed negative relationship
between "excess" profits and "welfare losses" was explainable on the
grounds that our estimates show the amount by which consumer "welfare"
would increase if that industry acquired or divested itself of the
appropriate amount of resources. Hence, we cannot interpret large
"welfare losses" in a given industry as a sign that the industry is
highly monopolistic. It may be that it is highly competitive. The
more relevant figure for ranking of industries by monopoly power is
Lerner's index. We have computed and shown such a ranking. Also
of note is the negative correlation between absolute and relative
"welfare losses" and the size of the total capital base. The impli-
cation that the smaller the industry, in asset terms, the larger
the "welfare losses" is certainly interesting. This may be partially
explained by the fact that larger firms may be able to hide "quasi—
monopoly" elements and may have economies of scale.
We estimated average two-digit concentration ratios (CR)
by utilizing available four—digit information. This was done for
value-added and employment data for the 4, 8, 20 and 50 largest
118
firms: We needed the two-digit estimates since most of our data
1mm: in this form. (Of course, the intercorrelations among the CR
rane quite high--.90 being the lowest coefficient). Since these
CR are presumed to reflect the degree of monopoly power in an
industry_we wanted to correlate these CR with a number of our
findings.
Correlations of CR with fully adjusted data, whether they be
for profit rates, welfare losses or Lerner's index, yielded higher
results than unadjusted data. This is in line with our a priori
expectations and justify our adjustment process--assuming the CR
are accurate indices of the real monopoly power. High correlation“
of Lerner's index and CR was perhaps the more significant result
since high "welfare losses" per se in a particular industry are
not necessarily indicative of high concentration in that industry.
It is also important to note that since both CR and Lerner's index
purport to measure percentage deviations, their relationship is
more significant than CR on, say, absolute "welfare losses." On
the negative side, it should be mentioned that the marginal elas-
ticity estimates did not always yield higher correlation coefficients
on CR than the unity elasticity estimates. However, for most of the
findings the expected results were obtained.
So as to not get lost amongst the wealth of secondary and even
peripheral findings and conclusions, we want to reiterate that the
most significant result, as far as the hypothesis we are testing here,
is that the estimated "welfare losses" in the American economy are
119
of a significantly different and higher order of magnitude than
previous studies had indicated.
Using what appears to be the most realistic estimates--based
on after-tax income, fully adjusted with industry-by-industry
elasticity data--we obtain a "welfare loss" of roughly six per
cent of national income.
We say all of this subject to all the theoretical and statis-
tical qualifications we discussed at the endof Chapter IV. Any
interpretation or evaluation of our results or conclusions should
be done in light of these shortcomings. Without going into these
in any detail, we want to single out the imperfection of data
problem as being especially unfortunate. This imperfection
forces us to neglect certain elements which may be of a quite
important character. The bias which may result from the imper-
fect nature of the existing data cannot be predicted a priori
with any great accuracy. However, we do want to at least £2333—
1§£g_on the relative magnitudes of some of the factors we
neglected.
Although it is true in general that we tried to be more
moderate in our estimates than Harberger,who tried to overestimate
the losses on every occassion, on balance, we probably leaned
toward overestimation. However, we feel this tendency is more
than swamped by the fact that most of the more or less arbitrary
_adjustments which we neglected which would lead to underestimation
of the losses far surpass in importance the items we neglected that
would lead to overestimation of the loss. In particular, we feel
120
the underestimation caused by neglect of mergers, redistributional
effects and monopoly gains to other factors would greatly increase
our losses if it were possible to calculate them-~net, of course,
of the more important balancing items such as allocating but one-
third of profits to monopoly power (in Harberger's case the change
from an 100 per cent allocation to a one-third distribution to
the highest profit industries until exhausted changed his estimates
very little), constant costs and unity elasticity assumption. This
last consideration is also reduced when we recall that elasticity
only enters into our loss estimating equation as it is, while
"excess" profit rates enter by a square of itself. Of course, we
are not intimating that the magnitude of the bias resulting from
the imperfect nature of the exisiting data can be predicted with
any great precision a priori. However, we do feel that the
direction of the bias can be speculated upon, and as indicated,
we feel it is in the direction of underestimating our losses.1
In conclusion, we feel that the monopoly problem takes on
a rather different perspective in the light of the present study.
The problem of monopoly acquires aggregative significance in
addition to its importance in studying particular industries. In
short, monopoly does affect aggregative "welfare” in a significant
way through its effect on resource allocation.
1
In this paper, we have not concerned ourselves with the question of how
the reallocation would be carried out in practice. This is an interesting
but secondary question here. For a brief taste of the lump-sum tax--
subsidy, etc., methods of doing this see Joan Robinson, op. cit., Chapter
13, pp. 159-165.
BIBLIOGRAPHY
Books
Allen, R. G. D. Mathematical Analysis for Economists. London: Mac-
millan and Co., 1938.
AR Bain, Joe S. Industrial Organization. New York: John Wiley and Sons,
Inc., 1959.
Croxton, F. and Cowden, D. Applied General Statistics. New York:
Prentice-Hall, Inc., 1939.
Dean, Joel. Managerial Economics. Englewood Cliffs, N. J.: Prentice-
Hall, Inc., 1951.
Dixon, W. and Massey, F. Introduction to Statistical Analysis. New
York: McGraw-Hill Book Company, Inc., 1957.
; Epstein, Ralph C. Industrial Profitsfiin the United States. New York: \//
National Bureau of Economic Research, 1934.
Galbraith, John K. American Capitalism: The Concgpt of Countervailing
Power. Boston: Houghton Mifflin Company, 1952.
Henderson, James M. and Quandt, Richard E. Microeconomic Theory. New
York: McGraw-Hill Book Company, Inc., 1958.
Johnston, J. Econometric Methods. New York: McGraw-Hill Book Company,
Inc., 1963.
Kaplan, A. D. H. Bingnterprise in a Competitive System. Washington,
D. C.: Brookings Institute, 1954.
\ierner, Abba P. The Economics of Control. New York: Macmillan Company,
' 1944.
Liebhafsky, H. H. The Nature of Price Theory. Homewood, Illinois: \/
The Dorsey Press, Inc., 1963.
Little, I. M. D. A Critique of Welfare Economics. Second edition. \/
London: Oxford University Press, 1950.
.:\Mach1up, Fritz. The Political Economy of Monopoly. Baltimore: The \/
" John Hopkins Press, 1952.
121
122
Malanos, George. Intermediate Economic Theory. Chicago: J. B. Lippin-
cott Company, 1962.
Meade, James E. Trade and Welfare. London: Oxford University Press,
1955.
Reder, Melvin. Studies in the Theory of Welfare Economics. New York:
Columbia University Press, 1947.
Roberts, David R. Executive Compensation. Glencoe, Illinois: The
Free Press, 1959.
Robinson, Claude. Understanding Profits. Princeton: D. Van Nostrand J
Cbmpany, Inc., 1961.
Robinson, Joan. The Economics of Imperfect Competition. London: Mac- J
millan Company, 1933.
Stigler, George J. ggpital and Rates of Return in ManufacturingyIndus- ~/
tries. New York: National Bureau of Economic Research, 1963.
Walker, H. and Lev, J. Statistical Inference. New York: Holt, Rinehart
and Winston, 1953.
Weiss, Leonard W.‘ Economics and American Industry. New York: John ¢
Wiley and Sons, Inc., 1961.
Weston, J. Fred. The Role of Mergers in the Growth of Large Firms.
Berkeley: University of California Press, 1953.
Articles and Periodicals
Adams, Walter. "Consumer Needs and Consumer Sovereignty in the
American Economy," The Journal of Business (July, 1962), pp.
264-277.
Aitken, A. C. "On Least-Squares and Linear Combination of Observations,“
Proceediggs of the Royal Society of Edinburgh, Vol. 55 (1934-1935),
pp. 42-48.
Bain, Joe S. "The Profit Rate as a Measure of Monopoly Power," V"
,Quarterly Journal of Economics, LV (1940-1941), pp. 271-293.
"Relation of Profit Rate to Industry Concentration: American
.Manufacturing, 1936-1940,"_Quarterly Journal of Economics, LXV
(August, 1951), pp. 293-324. '
Dorfman, Robert and Steiner, Peter. "Optimal Advertising and Optimal
Quality," American Economic Review, XLIV (December, 1954),
123
Economic Research Department, Chase Manhattan Bank (New York. Business
in Brief (November-December, 1963), pp. 4-5.
Garbarino, Joseph W. "A Theory of Interindustry Wage Structure Varia-
tion," Quarterly Journal of Economics, LVIV (May, 1950), pp. 285-
305.
Harberger, Arnold C. "Monopoly and Resource Allocation," Proceedingg
of the American Economic Review, XLIV (May, 1954), pp. 77-87.
"The Measurement of Waste,” Proceedings of the American
Economic Review, LIV (May, 1964), pp. 58-76.
Hotelling, Harold. "The General Welfare in Relation to Problems of
Taxation and of Railway and Utility Rates," Econometrica, VI,
no. 3 (July, 1938), pp. 242-269.
5Lerner, Abba P. "Monopoly and the Measurement of Monopoly Power,"
Review of Economic Studies, I (June, 1934), pp. 157-175.
Lipsey, R. G. and Lancaster, Kelvin. ”The General Theory of Second
Best," Review of Economic Studies, Vol. 24 (1956-1957), pp. 11-32.
Mack, Ruth P. "Discussion," Proceedings of the AmericanyEconomic
Review, XLIV (May, 1954), pp. 88-89.
McGurie, Joseph W., Chiu, John S. Y., and Elbing, Alvar 0. "Executive
Incomes, Sales, and Profits,” American Economic Review, LII
(September, 1962), pp. 753-761.
Schwartzman, David. "The Effects of Monopoly on Price," gpurnal of
Political Economy, Vol. 67 (August, 1959), pp. 352-362.
"The Burden of Monopoly," Journal of Political Economy,
Vol. 68 (December, 1960), pp. 627-630.
"The Effect of Monopoly: A Correlation,” Journal of
Political Economy, Vol. 69 (October, 1961), p. 494.
”The Economics of Antitrust Policy," The Antitrust Bulletin,
VI, no. 3 (May-June, 1961), pp. 235-243.
"Monopoly and Wages," Canadian Journal of Political Economy,
Vol. 26 (August, 1960), pp. 428-438.
Stigler, George J. "The Statistics of Monopoly and Merger," ggprnal of
Political Economy, Vol. 44 (February, 1956), pp. 33-40.
Telser, Lester G. "How Much Does It Pay Whom to Advertise," Proceedings
of the American Economic Review, LI (May, 1961), pp. 194-205.
124
Zellner, Arnold. "An Efficient Method of Estimating Seemingly
Unrelatelee8r98810n81nd Test for Aggregation Bias," Journal of
the American Statistical Association, Vol. 57 (June, 1962), pp.
348-368.
K "Estimators for Seemingly Unrelated Regression Equations:
Some Exact Finite Sample Results," Journal of the American
Statistical Association, Vol. 58 (December, 1963), pp. all-992.
Public Documents
U. S. Bureau of the Census. Concentration Ratios in Manufacturing
Industry, 1958. Parts I and II. Washington, D. C.: Government
Printing Office, 1963, 1962.
U. S. Treasury Department. Internal Revenue Service, Statistics of
Income, U. 8. Business Tax Returns, 1957-1958, 1958-1959, 1959-
1960, 1960-1961.
U. S. Treasury Department. Internal Revenue Service, Statistics of
Income, U. S. Corporation Tax Returns, 1956-1957, 1957-1958,
1958-1959, 1959-1960, 1960-1961.
McConnell, Joseph J. "1942 Corporate Profits by Size of Firms,"
Survey of Current Business (January, 1963).
APPENDICES
125
APPENDIX A
MEASUREMENT AND ESTIMATION PROCEDURES
126
127
MEASUREMENT AND ESTIMATION PROCEDURES
In our analysis of the ”welfare losses" we used a number of
variables--some of which could only be estimated by rather crude
methods. Our chief source of data was the comprehensive lfigL
Statistics of Income for Corporations (C) for 1956-57 to 1960-61
and:§usiness [i.e“,partnerships (P) and sole proprietorship (SP)
in addition to C] for 1957—58 to 1960-61. The measurement and
estimation procudeures employed on the relevant variables is
discussed in the paragraphs below.
Industrial Classification
Since the industrial classification employed by 135 (basically
a two-digit standard industrial classification) changed over the
years studied and the fact that P and SP have different formats
from C, made it necessary to reduce all three types of business
establishments to a common denominator. We used the 1960-61 C
industry proportions as our benchmark. When we could not establish
the proper proportions for P and SP from other year's information
we utilized the benchmark proportions.
Net Income
In computing profit rates, we used two different income
figures for Co-before-and-after corporation income taxes. This
figure was then combined with the untaxed income figures of P
and SP. The figures for all types of business were then more
unifomm in that they all neglected personal income taxes.
128
Capital Accounts
It was necessary to estimate capital figures for P and SP
as it was for most of the balance sheet information on P and SP.
Income statements were,in general, much more available and probably
more reliable.
We utilized two different capital bases in computing profit
rates,yig,, equity (net worth or capitalization) and total capital.
Under equity we included preferred and common stock, paid-in or
capital surplus, surplus reserves and earned surplus and undivided
profits. Total capital was computed by adding funded debt (capital
borrowed from the general public and lending institutions through
the sale of bondS, debentures and other forms of indebtedness--
specifically, for IRS data, we included bonds, notes, mortgages
of any length). Of course, in computing profit rates we used
total profit,i.e., net income (net earnings after all business
expenses and fixed charges including interest payments on funded
debt have been deducted) plus interest payments on funded debt as
the relevant income figure when we used total capital. The reason
that the returns are computed to include funded debt is that these
borrowed dollars perform.much the same economic function as invested
capital. In general, the return on total capital will be lower for
most companies since earnings usually exceed the interest charged
to the firm.
On both types of capital figures, we made adjustments since
the data are shown for end-of-year assets rather than average (or
possibly mid-year) assets. The difficulty is that when the rate of
129
growth of assets is very high, the rate of return is seriously
underestimated. Although a constant geometric rate of increase
of assets might be more plausible, we computed the return on a
simple linear assumption in the following manner: Letting Ao
be the assets at the beginning, and A assets at the end of the
1
year and R be income, the average rate (AR) is AR:= R/[1/2(Ad+Alfl .
The capital figures for P and SP were estimated from two
sources: (1) from the equity and total capital to business re—
ceipts (gross sales plus gross receipts from operations) ratios
found in the available balance sheet data for P for the 1959-60
period; (2) from the total capital to business receipts ratio in
small corporations for each of the years investigated.
The PfiBalance Sheet Approach
In this approach, we were forced to make some rather strong
assumptions. First of all, balance sheet data were reported by
only 44.5 per cent of the P filing income statementse-although in
some industries the percentages exceeded 90. So we have to assume
the percentage not reporting in each industry had the same ratios
as the reporting firms and "blow-up" our figures to 100 per cent
in this manner. Thus, not only did we have to assume this period
was representative of the other three years in P, but, we had to
assume the same proportions applied to P and SP. Remember, how-
ever, that since it is the combined capital figures of C, P and
SP that we are really interested in and since the known corporate
sector is the largest component in most sectors, this crudeness
can be somewhat justified.
130
The Small Corporation Balance Sheet;§ppgoach
As an alternative formulation, we have adopted the approach
used by George J. Stigler in his Capital and Rates of Return in
Manufacturing:Industries, (National Bureau of Economic Research, 1963),
pp. 7, 8, 114-118, 221, to estimate the capital of noncorporate
enterprises--assuming annual industry data on sales (or business
receipts) are available, as is the case. His estimate being based
upon theratio of capital to sales in small C assumes that noncor-
porated enterprises more closely resemble small rather than all C.
It is felt that it would be undesirable simply to use the ratio
found in the entire corporate sector because: (a) most noncorporate
enterprises are small; (b) small corporations typically have rela-
tively low ratios of capital to receipts-~in fact, Stigler found
that the ratio was almost twice as large in the asset class over
$100 million as it was in the under $50,000 class. We computed
the total capital to receipts ratio for the $O-25,000 total asset
class and the total assets to receipts ratio for the $0-50,000 total
asset class on an industry-by-industry basis for each of the four
years studied. Since the former is the first enumerated class it
may contain too much of a "catch-all" (residual) character.. Since
the ratios are significantly different, the latter is probably the
more useful (we base part of this on the fact that in our brief
survey this smaller asset class had what appeared to be "unreason-
able" profit rates). Remember, however, although the larger the
noncorporate sector the more unreliable the figure, the corporate
sector usually dominates. Indeed, Stigler found only one three-
digit industry dominated by the noncorporate sector-~the fur-garment
131
industry, see ibid., p. 117. (Incidentally, no estimate of equity
was made under this approach, but, we did in (1) by finding equity
to total capital ratio in P).
Interest Paid
Although this information was complete for C, we had to
estimate it for P and SP. Fortunately, for the 1959-60 period
we had data for both P and SP. So we merely took the industry-
by-industry ratio of interest paid to average total capital and
made the proper multiplication to obtain our estimates. Of course,
we have to again assume this period was representative.
Intangible Assets and Royalties
Here again the information on C was complete, but we had to
estimate for P and SP assuming the industry-by-industry intangible
assets to average total capital and royalties to net income ratios
for 1959-60 period were representative and applicable to P and SP
for the entire four-year period investigated.
Profit Rate and "Welfare Loss" Adjustments
" we made
In computing profit rates and hence "welfare losses,
adjustments for intangibles, royalties and advertising. In each
case we made the somewhat arbitrary assumption that each of these
elements was a 100 per cent monopoly element. Therefore, in com-
puting profit rates,our estimates became larger and larger as we
made these cumulative adjustments. In using this information for
estimating "welfare losses" we used a slightly different procedure.
132
We subtracted intangibles from the capital base and added advertising
to the net income figures of each industry. This increases the indus-
try profit rates. In the case of royalties, we subtracted them from
the overall average or "normal" profit rate-~a1ternatively, we could
have employed the same method we utilized for advertising expenditures,
but one was easier for computational purposes. When we came to esti-
mating "welfare losses," we ran into some trouble as raising the pro-
fit rate of an industry earning less than "normal" profits meant the
profit figure became larger as it became a smaller negative number;
but, it declined in absolute value. And since the losses involve
squaring "excess" profit rates our losses became smaller in those indus-
tries after adjustments--in fact, in some cases, they overcame the
plus items resulting in a lower estimate adjusted than unadjusted!
We got around this by merely reducing the industry profit rate by
the corresponding difference in the negative cases,i.e., increasing
the absolute value of the losses since a lower industry profit rate
subtracted from a constant overall rate increases the "excess" profit
rate makes a greater differential when squared. The profit rates
found by making the minus adjustment for below average profit indus-
tries, we have numbered Methods I-VIII; while the other method,
increasing the profit rates by the adjustment in all cases, as
Harberger did, we have labeled Methods 1-8. In general, we used
Methods I-VIII only for computing "welfare losses." Therefore, our
tables of frequency distributions, etc., on profit rates utilize
Methods 1-8 (in the raw form, one can easily check which method is
being employed by observing whether the first-to-last column is
133
greater than or less than the preceding column--if it is greater
Methods 1-8 are being utilized, if less Methods I-VIII).
AdvertisiggyExpenditures
Although the information on C is bountiful, it is non-
existent for P and SP. Our use of the C advertising to business
receipts ratios may be bias since, e.g., in retailing, which is
more important in SP and P than C, "wastesful" advertising is
less significant than in manufacturing, which is a quite important
segment of C.
APPENDIX B
PROFIT RATE DATA FOR THE AMERICAN ECONOMY
1956-1957 TO 1960-1961
134
135
1? Lb \x
mmm.eH omm.~H mm~.a Hse.s mmm.~H mHo.HH omm.m mHo.a can:
as as as co as on as as Hausa
a o o o o o o o Ho>o was oo.s~
a o o o o o o o mm.m~ - oo.s~
N N H o m o o o ma.m~ - oo.-
m H o o m H o o sm.H~ - oo.o~
m a N o m N o o am.aH - oo.mH
oH s a o a N n H mm.HH - oo.sH
a m s o a a H o mm.mH - oo.sH
m a a ~ a a m o aa.nH - oo.NH
s a HH m m 0H s m mm.HH - oo.oH
m s 0H s a HH SH a mm.m - co.»
m s m 0H s m a mH ma.a - 00.0
N a a HH s NH oH SH ma.m - oo.s
o H s w o o m NH am.m - oo.~
H H H m H H H H sm.H - oo.o
@ 0.“ Nu mnmvCH NO H0952 mm.“ .Hu mafia HO Hflflafiz
A3 A3 A3 A3 A3 a: £3 a. 93 3V
.HHV tom euumsfiesen
amusemea
meoouH xmfi
cowumuomuoonouomum
uaoooH any
sowumuomuoosuuum<
maoocH xma
soaumuomuoouquMQm
oaoocH xma
oowumuomuoosuoumm
omen magnum Aummhumaz uov owmum>< mean:
omen Hmuqmmo Hmuoa Aummaumwz uov oweuo>< mafia:
manna mwmucoouom
ca oumm uwuoum
Homau coma Ou mmma omma mofluum
umowno>ah onu How saunas“ mwmum>< son: momma esowumuonuou no nouns mamoum mo sowuonwuueaa hosmsvoum
II I“
*mmHmmMOHMHmmoxm MAOm 92¢..mmHmmmMZHM¢N .mchadzommoo mom mdem HHmdmm mu¢mm>d.m0 mZOHHDmHmHmHQ wuzmadmmh
Hum mamdfi
136
mmo.mN omm.mN oam.0N coo.mH mqa.qa oeN.NH use:
co co co co co 00 Nancy
mH w NH o e m uu>o mum oo.m¢
0H mm m m o N am.N¢ . oo.mm
m N m a N N mm.Nm u oo.om
m m m e x e N mm.mN . oo.NN
N n a N N m mm.oN u oo.¢N
N m m N m N mm.mN u oo.HN
N o m H m m mm.0N u oo.mH
m N m m m m mm.NH u oo.mH
m m N N 0 0H mm.¢N a oo.NH
H m e m H m mm.HH u oo.m
m m o w o N mm.» 1 00.0
o o a m w m am.m u oo.m
N H e n m m mm.N u oo.o
mumumnmsH we uoaanz moauumovoa mo umnenz
«ea AHV Nev Nev A< omen huwovu omen HmuNamo Hmuoa cw comm uawoum
Ammomuvfiz soy mmmum><
Ammuhuvax HOV ammuo><
HeHHaao uuaaHumm cu moauae
Human museamm mooaueuoouoo Hausa moan:
HauHaau uueaHHnm
Ou moaned amonm musmasm afinmucsuumm moan:
uaoocH mecuuum<
Homaaooma Ou mmmHINan .mowuum
umowrunom emu now mauauum emeuo>< com: momma mmwumuoouumm mo mound uNmoum we soauaaauuman hoooovoum
AeoseHuaoov--H-m mamas
137
.wumo% ucm>oHou can now
monoumm me oaoooH mmosHuam mom .mcuouom xma msoocH coHuwuomuou .osoosH mo onumHuwum mmH. "momsom
.wousqucmmxo mchHuuo>vm n Adv mmoHuHmhou n Amy “museum oHnchmuuH n AHVRR
=.muooeumomm< .eeoH euemHoZ.
use comm uHmoum: cOHmmoomHv .< vaaommd mom “HHH>1H mos .wuH sponge: How one momma uHmoum mmmnHe
«em.Nm MNN.mm NoH.NN Nnm.mH wem.¢H omm.MH one:
om O© oo co oo oo HmuoH
oH HH w e uo>o mam oo.me
HH m uo>o use oo.me m m m o mm.N¢ u oo.mm
H N mm.N¢ u oo.m¢
m H mm.¢¢ u oo.Ne H m m m mm.Nm 1 00.0m
N N mm.H¢ u oo.mm e N a N mm.mN u oo.NN
m e mm.wm u oo.om N H m a mm.oN u oo.¢N
q n mm.mm u oo.nm m m a m mm.mN u oo.HN
H e mm.Nm 1 oo.on H m m m mm.0N u oo.mH
m c m w mm.NH u oo.mH
m N mm.mN u oo.¢N m N m N mm.¢H a oo.NH
o NH mm.nN 1 oo.mH n e N m mm.HH u oo.m
N o mm.NH a oo.NH n q o o mm.w 1 oo.o
a 0 OH 0 mo.m u oo.n
o a mm.HH 1 oo.o e e e o mm.N n 00.0
moHuumouoH mo nonesz moHuumomuH mo nooaoz
3 8V 3 8v 3 as
mean HmuHmeo Hmuoa Aum051VHz uOV omeno>< omen huHovm use» omen Hmanmo Hmuoe
.anz HOV uwmuo>< AummwumHz HOV owmuo>< magma ammusoouom
sH comm unoum
HeuHemo oueaHuam ou HauHemo someHumm ou
mOHumm umonm oocmHmm cOHumuomHoo HHmam wchD v moHumm uuozm moomHmm mHnmuosuumm wch:
oaoocH xmanuoum<
HomHaocaH ou wmmHuNmmH .vOHuom umowunsom
mnu now mcuouom ammue>< com: momma mmHseuouuHumoum eHom Ho mound unoum mo soHunnHuumHa hosoovoum
«a..ulll|=aHs=oov--H-HH Sea
138
TABLE B-2
AVERAGE PROFIT RATES IN CORPORATIONS USING BEFORE- AND AFTER-
TAX INCOME AND USING EQUITY AND TOTAL CAPITAL BASES
1956-1957 to 1960-1961
=*_._ I m
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles (I) Royalties(R) Advertising(A)
After-Corporation Tax Income--Average (or Mid-year) Equity Base
V'—
1 .07013 .07057 .07230 .08859
2 .03076 .03086 .03383 .04097
3 .03718 .03788 .04272 .04415
4 .02785 .02870 .03018 .03032
5 .02452 .02458 .02637 .02728
6 .04187 .04258 .05128 .05282
7 .04781 .04889 .05083 .05473
8 .05743 .05759 .05814 .06711
9 .06909 .06977 .07193 .09670
10 .06481 .06556 .06690 .16309
11 .07076 .07105 .07199 .13952
12 .09518 .09523 .09587 .16491
13 .04605 .04611 .04755 .05931
14 .05400 .05418 .05728 . ..Q9528
15 .06464 .06480 .06625 .07419
16 .06050 .06067 .06115 .09659
17 .07451 .07469 .07553 .08655
18 .08807 .09020 .09473 .10913
19 .09548 .09607 .10118 .15256
20 .03859 .03997 .04209 .04779
21 .07325 .07345 .07431 .10690
22 .05803 .05811 .05857 .08926
23 .07959 .07977 .08158 .09232
24 .06472 .06479 .06557 .07087
25 .06570 .06594 .06704 .08671
26 .07227 .07268 .07671 .09374
27 .08029 .08066 .08434 .12251
28 .07098 .07124 .07379 .08156
29 .10448 .10498 .10572 .12177
30 .09172 .09250 .09528 .13633
31 .05751 .05787 .05976 .08797
32 .04301 .04309 .04340 .04549
33 .02860 .02867 .02912 .03226
34 .06443 .06469 .06505 .06819
35 .04521 .04522 .04541 .04620
36 .04171 .04188 .04190 .04218
37 .06731 .06761 .06845 .11918
38 .06280 .06302 .06428 .08981
39 .06021 .06048 .06132 .08959
139
TABLE B-2--(Continued)
m m M
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles (I) Royalties(R) Advertising(A)
40 .06142 .06160 .06313 .08961
41 .06438 .06460 .06591 .09137
42 .05410 .05428 .05447 .10982
43 .06288 .06307 .06332 .10637
44 .07014 .07022 .07027 .13117
45 .04220 .04234 .04242 .11632
46 .02705 .02710 .02719 .10831
47 .03115 .03120 .03125 .07801
48 .03894 .03985 .04102 .07774
49 .03801 .03806 .03826 .05775
50 .05549 .05590 .05628 .10052
51 .05216 .05244 .05300 .08672
52 .11401 .11442 .11691 .12101
53 .17038 .17155 .17170 .17793
54 .09626 .09635 .09638 .10016
55 .08124 .08141 .08904 .08920
56 .05859 .05870 .05903 .06632
57 .11986 .11997 .12020 .12252
58 .10689 .11111 .11132 .13265
‘ 59 .04475 .04489 .04510 .04815
60 .03591 .03604 .08464 .08487
61 .05450 .05557 .05896 .08603
62 .03040 .03051 .03070 .04883
63 .06610 .06709 .06768 .10797
64 .08469 .08605 .09387 .11391
65 .05873 .05914 .05948 .07386
66 .03108 .03294 .03873 .07874
67 .04952 .05036 .05319 .08803
68 .08219 .08375 .08675 .12958
69 -.OO318 -.00323 -.00206 .00300
After-Corporation Tax Income-~Average (or Mid-Year) Total Capital
Base
1 .06691 .06750 .06994 .09288
2 .02352 .02363 .02815 .03903
3 .03525 .03612 .04249 .04438
4 .02353 .02452 .02654 .02674
5 .01986 .01992 .02217 .02331
6 .04444 .04548 .05730 .05939
7 .04704 .04837 .05078 .05564
8 .06252 .06278 .06360 .07694
9 .07473 .07564 .07836 .10948
10 .07031 .07139 .07314 .19914
11 .07602 .07642 .07763 .16475
140
TABLE B-2--(Continued)
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles (I) Royalties(R) Advertising(A)
12 .12678 .12687 .12784 .23245
13 .04378 .04385 .04562 .06018
14 .05011 .05033 .05446 .10489
15 .05442 .05457 .05613 .06467
16 .06096 .06117 .06176 .10533
17 .08407 .08431 .08540 .09955
18 .09600 .09883 .10439 .12206
19 .11642 .11739 .12424 .19316
20 .03893 .04058 .04312 .04992
21 .08497 .08528 .08646 .13116
22 .05671 .05680 .05739 .09649
23 .08537 .08559 .08774 .10056
24 .07042 .07052 .07154 .07847
25 .06834 .06865 .07000 .09429
26 .07731 .07786 .08289 .10414
27 .09072 .09127 .09609 .14610
28 .07878 .07919 .08275 .09362
29 .11176 .11237 .11321 .13136
30 .10405 .10520 .10881 .16219
31 .05854 .05902 .06145 .09773
32 .04677 .04692 .04748 .05121
33 .02156 .02165 .02238 .02746
34 .07888 .07936 .07991 .08479
35 .05410 .05412 .05453 .05620
36 .04489 .04527 .04531 .04591
37 .05569 .05596 .05691 .11377
38 .06834 .06868 .07048 .10690
39 .06511 .06551 .06669 .10621
40 .06423 .06448 .06651 .10162
41 .06734 .06767 .06947 .10458
42 .05458 .05483 .05508 .13028
43 .09851 .09906 .09951 .17719
44 .07359 .07369 .07376 .14814
45 .04114 .04131 .04142 .13504
46 .01593 .01597 .01609 .12973
47 .01229 .01232 .01240 .09093
48 .03762 .03917 .04122 .10553
49 .03130 .03135 .03159 .05570
50 .05715 .05774 .05826 .11938
51 .05304 .05343 .05418 .09955
52 .07677 .07714 .08053 .08610
53 .07014 .07064 .07078 .07720
54 .04190 .04202 .04212 .05444
55 .08473 .08492 .09314 .09331
56 .09146 .09196 .09297 .11516
141
TABLE B-2--(Continued)
w m
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles (I) Royalties(R) Advertising(A)
57 .11909 .11920 .11943 .12177
58 .12490 .13122 .13149 .15882
59 .04790 .04834 .04898 .05810
60 .03766 .03791 .12793 .12835
61 .06149 .06373 .06999 .11994
62 .00892 .00900 .00947 .05422
63 .07383 .07544 .07630 .13486
64 .10539 .10801 .12023 .15151
65 .07047 .07169 .07252 .10792
66 .02580 .02852 .03876 .10950
67 .05084 .05254 .05815 .12707
68 .09672 .09935 .10363 .16477
69 -.01833 ~.01887 -.Ol69l -.00849
Before-Corporation Tax Income--Average (or Mid-year)Tota1 Capital
Base .
1 .11019 .11088 .11261 .12889
2 .05445 .05462 .05759 .06473
3 .07886 .08034 .08518 .08661
4 .07496 .07726 .07873 .07888
5 .04017 .04026 .04205 .04296
6 .09268 .09427 .10297 .10451
7 .08036 .08219 .08412 .08802
8 .10478 .10507 .10562 .11459
9 .12902 .13028 .13244 . .15721
10 .12021 .12161 .12295 .21914
11 .13233 .13288 .13381 .20134
12 .16234 .16241 .16306 .23210
'13 .08288 .08299 .08442 .09618
14 .09532 .09564 .09875 .13674
15 . .09826 .09851 .09996 .10790
16 .12484 .12520 .12568 .16111
17 .13691 .13722 .13807 .14909
18 .16120 .16509 .16962 .18402
19 .14594 .14685 .15196 _.20334
20 .05900 .06110 .06322 .06892
21 .14117 .14154 .14240 .17499
22 .10801 .10816 .10862 .13931
23 .15353 .15387 .15567 .16642
24 .11967 .11981 .12058 .12588
25 .12714 .12760 .12870 .14837
26 .14102 .14182 .14585 .16288
27 .15952 .16026 .16395 .20212
28 .14667 .14721 .14976 .15753
142
TABLE B-2--(Continued)
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles (I) Royalties(R) Avertising(A)
29 .20935 .21036 .21110 .22715
30 .18083 .18236 .18514 .22620
31 .09934 .09996 .10186 .13006
32 .07174 .07187 .07218 .07427
33 .04487 .04498 .04543 .04857
34 .11838 .11885 .11920 .12235
35 .07314 .07315 .07334 .07413
36 .05862 .05885 .05887 .05916
37 .12176 .12229 .12314 .17387
38 .10264 .10299 .10426 .12979
39 .11649 .11700 .11784 .14611
40 .11608 .11642 .11796 .14444
41 .10076 .10111 .10242 .12789
42 .09618 .09650 .09668 .15204
43 .10640 .10673 .10697 .15002
44 .13206 .13222 .13227 .19317
45 .07289 .07314 .07322 .14712
46 .04690 .04698 .04707 .12819
47 .05050 .05058 .05063 .09738
48 .07563 .07741 .07858 .11530
49 .05603 .05611 .05630 .07579
50 .08470 .08533 .08571 .12995
51 .08433 .08478 .08534 .11906
52 .13079 .13126 .13375 .13785
53 .19674 .19809 .19824 .20447
54 .10674 .10684 .10687 .11065
55 .08609 .08627 .09390 .09406
56 .07829 .07843 .07876 .08604
57 .13733 .13746 .13768 .14001
58 .15178 .15778 .15799 .17932
59 .05552 .05569 .05591 .05896
60 .05149 .05167 .10028 .10050
61 .09031 .09209 .09548 .12254
62 .04748 .04765 .04784 .06597
63 .10038 .10189 .10248 .14276
64 .14378 .14607 .15390 .17394
65 .08063 .08119 .08153 .09591
66 .05687 .06026 .06605 .10606
67 .09473 .09633 .09916 .13400
68 .14421 .14695 .14995 .19278
69 .00768 .00782 .00899 .01405
143
TABLE B-2--(Continued)
Adjusted for
Industry Unadjusted Adjusted for (I) and
Adjusted for
(I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
Before-Corporation Tax Income Average (or Mid-Year) Equity Base
.12320 .12428 .12673 .14967
1
2 .05951 .05979 .06432 .07519
3 .08979 .09201 .09838 .10027
4 .08746 .09114 .09316 .09336
5 .03950 .03962 .04186 .04300
6 .11307 .11570 .12752 .12961
7 .08741 .08988 .09230 .09716
8 .13282 .13336 .13418 .14752
9 .14983 .15166 .15438 .18549
10 .14263 .14481 .14656 .27256
11 .15536 .15619 .15740 .24452
12 .22850 .22866 .22964 .33424
13 .08935 .08950 .09127 .10583
14 .10491 .10538 .10950 .15994
15 .09061 .09086 .09242 .10096
16 .14001 .14050 .14109 .18465
17 .16413 .16461 .16570 .17985
18 .18523 .19070 .19626 .21393
19 .18396 .18550 .19235 .26127
20 .06314 .06582 .06835 .07516
21 .17802 .17867 .17985 .22455
22 .12037 .12057 .12116 .16026
23 .17352 .17397 .17613 .18895
24 .14225 .14246 .14348 .15041
25 .14414 .14479 .14614 .17043
26 .16297 .16412 .16915 .19039
27 .19438 .19556 .20038 .25039
28 .18454 .18550 .18906 .19993
29 .23027 .23153 .23236 .25051
30 .21961 .22203 .22564 .27902
31 .11225 .11316 .11560 .15187
32 .09817 .09849 .09906 .10279
33 .04785 .04804 .04878 .05386
34 .16235 .16335 .16390 .16879
35 .11319 .11323 .11364 .11532
36 .08076 .08145 .08149 .08209
37 .11668 .11725 .11820 .17506
38 .12507 .12569 .12750 .16391
39 .14364 .14452 .14570 .18522
40 .13663 .13716 .13920 .17430
41 .11745 .11801 .11981 .15492
42 .11168 .11219 .11244 .18764
43 .17685 .17784 .17829 .25597
144
TABLE B-2--(Continued)
Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
44 .14921 .14942 .14949 .22388
45 .07999 .08033 .08043 .17406
46 .04371 .04382 .04394 .15758
47 .04474 .04486 .04494 .12347
48 .10078 .10495 .10700 .17130
49 .05358 .05367 .05391 .07801
50 .09739 .09840 .09892 .16004
51 .09626 .09696 .09771 .14308
52 .09956 .10005 .10343 .10901
53 .09726 .09795 .09809 .10451
54 .07601 .07624 .07633 .08866
55 .08995 .09016 .09837 .09854
56 .15121 .15203 .15305 .17523
57 .13666 .13679 .13702 .13935
58 .18180 .19099 .19127 .21859
59 .07992 .08064 .08128 .09040
60 .06643 .06686 .15689 .15731
61 .12652 .13112 .13738 .18734
62 .05084 .05129 .05176 .09651
63 .12332 .12601 .12688 .18543
64 .19682 .20172 .21394 .24522
65 .12383 .12597 .12680 .16220
66 .06950 .07683 .08708 .15782
67 .13885 .14348 .14909 .21801
68 .18456 .18956 .19385 .25498
69 -.OOO48 -.OOO49 .00146 .00988
*For Industry Number see coding Table C-l.
SOURCE:
See Table B-1.
145
TABLE B-3
AVERAGE PROFIT RATES FOR ALL BUSINESS ESTABLISHMENTS--PROFIT
METHODS 1, 2, 3, 5, 7, 8*
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
Using Profit Method 1
1 .08992 .09051 .09197 .10689
2 .05242 .05287 .05412 .05730
3 .02841 .02898 .03540 .03747
4 .03033 .03116 .03243 .03262
5 .02584 .02593 .02763 .02856
6 .02123 .02176 .03450 .03662
7 .08678 .08882 .09065 .09566
8 .18775 .18826 .18852 .20308
9 .07327 .07399 .07608 .10093
10 .06980 .07058 .07184 .16855
11 .07533 .07571 .07659 .14433
12 .09843 .09849 .09913 .16834
13 .04785 .04791 .04932 .06112
14 .08034 .08060 .08336 .12178
15 .07222 .07257 .07362 .08063
16 .08652 .08726 .08774 .12651
17 .07648 .07666 .07750 .08854
18 .10748 .10998 .11413 .12841
19 .09609 .09669 .10175 .15318
20 .03977 .04116 .04325 .04895
21 .07549 .07570 .07655 .10927
22 .06383 .06392 .06437 .09544
23 .08599 .08620 .08975 .10092
24 .06641 .06649 .06726 .07260
25 .07043 .07068 .07169 .09050
26 .07841 .07885 .08283 .09988
27 .08141 .08178 .08542 .12340
28 .19385 .19458 .19709 .20486
29 .10721 .10773 .10845 .12513
30 .09460 .09540 .09814 .13923
31 .05881 .05917 .06104 .08937
32 .04642 .04654 .04684 .04904
33 .03315 .03325 .03358 .03634
34 .06527 .06555 .06591 .06905
35 .04531 .04534 .04553 .04632
36 .04174 .04192 .04194 .04223
37 .12283 .12404 .12454 .16546
38 .10464 .10496 .10596 .13089
39 .12070 .12116 .12181 .15685
40 .07823 .07844 .07945 .09904
146
TABLE B-3--(Continued)
*7
~
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
41 .10611 .10645 .10752 .13310
42 .11755 .11823 .11838 .16324
43 .14718 .14819 .14833 .19140
44 .08291 .08303 .08313 .13456
45 .10412 .10444 .10449 .16770
46 .09372 .09425 .09431 .16232
47 .09337 .09369 .09377 .13011
48 .18804 .19214 .19267 .21871
49 .07688 .07697 .07708 .09376
50 .13630 .13804 .13820 .17072
51 .09971 .10009 .10049 .12813
52 .08916 .08951 .09135 .09519
53 .14634 .14719 .14731 .15270
54 .08658 .08668 .08671 .09030
55 .07657 .07673 .08385 .08402
56 .05644 .05696 .05722 .08142
57 .10941 .10952 .10975 .11212
58 .59659 .61361 .61373 .64140
59 .03725 .03741 .03779 .04150
60 .02101 .02105 .04900 .04912
61 .34218 .34765 .34920 .37398
62 .03384 .03404 .03420 .04586
63 .25276 .25551 .25567 .28147
64 .21747 .22069 .22670 .24529
65 .21007 .21111 .21129 .23145
66 .03184 .03319 .03736 .06992
67 .10530 .11369 .11608 .15934
68 1.25093 1.26767 1.26842 1.31414
69 .13891 .13982 .14026 .15135
Using Profit Method 2
1 .10949 .11037 .11216 .13035
2 .26045 .27194 .27837 .29473
3 .03462 .03547 .04333 .04586
4 .03261 .03357 .03495 .03514
5 .02757 .02767 .02949 .03048
6 .02921 .03022 .04791 .05087
7 .08681 .08885 .09068 .09569
8 .21523 .21589 .21620 .23290
9 .07420 .07494 .07706 .10222
10 .07133 .07214 .07343 .17229
11 .07609 .07647 .07736 .14578
12 .09833 .09839 .09903 .16817
147
TABLE B-3--(Continued)
‘-
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
13 .04829 .04836 .04978 .06169
14 .08250 .08278 .08561 .12507
15 .08924 .08976 . .09106 .09975
16 .07444 .07498 .07540 .10871
17 .07697 .07715 .07799 .08910
18 .11096 .11363 .11792 .13267
19 -O9649 .09709 .10217 .15381
20 .04009 .04151 .04362 .04936
21 .07567 .07587 .07673 .10952
22 .06390 .06399 .06445 .09556
23 .08449 .08469 .08819 .09915
24 .06647 .06654 .06732 .07266
25 .07425 .07453 .07559 .09543
26 .07915 .07960 .08361 .10082
27 .08196 .08234 .08600 .12424
28 .19239 .19311 .19561 .20332
29 .10745 .10797 .10870 .12541
30 .09495 .09575 .09851 .13974
31 .05895 .05932 .06119 .08959
32 .04753 .04765 .04796 .05021
33 .04276 .04293 .04337 .04692
34 .06581 .06610 .06645 .06962
35 .04570 .04572 .04592 .04671
36 .04204 .04223 .04225 .04253
37 .14438 .14606 .14665 .19484
38 .10762 .10797 .10900 .13464
39 .12148 .12195 .12260 .15787
40 .09025 .09052 .09169 .11429
41 .10774 .10810 .10918 .13515
42 .14416 .14519 .14537 .20046
43 .16899 .17032 .17048 .21997
44 .09370 .09385 .09397 .15210
45 .11697 .11736 .11742 .18845‘
46 .12753 .12853 .12861 .22135
47 .12299 .12354 .12365 .17157
48 .18063 .18441 .18492 .20991
49 .09118 .09132 .09144 .11123
50 .18225 .18536 .18558 .22925
51 .11963 .12018 .12067 .15385
52 .11150 .11204 .11435 .11916
53 .17448 .17569 .17584 .18227
54 .09409 .09421 .09424 .09814
55 .08443 .08462 .09246 .09265
56 .13104 .13388 .13450 .19138
148
TABLE B-3--(Continued)
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
57 .12165 .12178 .12204 .12467
58 .42972 .43848 .43856 .45834
59 .04965 .04993 .05043 .05539
60 .03807 .03821 .08893 .08915
61 .49581 .50737 .50964 .54581
62 .07046 .07134 .07167 .09611
63 .42530 .43314 .43340 .47714
64 .26290 .26763 .27490 .29745
65 .31455 .31690 .31717 .34743
66 .04185 .04422 .04978 .09315
67 .11677 .12646 .12912 .17724
68 .43283 .45484 1.45570 1.50817
69 .11374 .11435 .11470 .12377
Using Profit Method 3
1 .10485 .10586 .10798 .12957
2 .07275 .07365 .07546 .08004
3 .02574 .02644 .03510 .03789
4 .02905 .03017 .03198 .03224
5 .02251 .02260 .02477 .02595
6 .01776 .01836 .03541 .03826
7 .05488 .05652 .05885 .06520
8 .28434 .28559 .28602 .30972
9 .07989 .08088 .08350 .11469
10 .07705 .07817 .07982 .20634
11 .08207 .08260 .08374 .17115
12 .13161 .13172 .13269 .23747
13 .04619 .04627 .04801 .06263
14 .08596 .08634 .09000 .14111
15 .06852 .06889 .07007 .07797
16 .09212 .09309 .09368 .14140
17 .08664 .08690 .08797 .10215
18 .12004 .12347 .23857 .14610
19 .11711 .11808 .12486 .19375
20 .03974 .04140 .04389 .05069
21 .08805 .08837 .08954 .13439
22 .06410 .06422 .06479 .10432
23 .09263 .09289 .09713 .11044
24 .07263 .07275 .07376 .08073
25 .07454 .07487 .07611 .09925
26 .08500 .08560 .09055 .11178
27 .09223 .09279 .09755 .14728
28 .23365 .23481 .23809 .24822
29 .11536 .11599 .11681 .13565
149
TABLE B-3-~(Continued)
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I),(R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
30 .10782 .10901 .11258 .16597
31 .06064 .06113 .06354 .09998
32 .05554 .05579 .05634 .06031
33 .04362 .04390 .04459 .05023
34 .07929 .07983 .08038 .08523
35 .05452 .05458 .05499 .05665
36 .04514 .04556 .04560 .04620
37 .13108 .13262 .13321 .18204
38 .12837 .12893 .13034 .16545
39 .14666 .14742 .14830 .19617
40 .08773 .08802 .08932 .11446
41 .12631 .12687 .12834 .16354
42 .13974 .14080 .14099 .19983
43 .21015 .21234 .21255 .27790
44 .08968 .08983 .08996 .15239
45 .11771 .11815 .11821 .19585
46 .10688 .10766 .10774 .19543
47 .11830 .11892 .11905 .17435
48 .22667 .23308 .23377 .26754
49 .08009 .08021 .08034 .10046
50 .16064 .16322 .16342 .20446
51 .11316 .11371 .11422 .14933
52 .07801 .07848 .08133 .08726
53 .06893 .06938 .06953 .07563
54 .03552 .03564 .03571 .04575
55 .08094 .08112 .08890 .08909
56 .11702 .11996 .12069 .18677
57 .11372 .11384 .11408 .11658
58 .83535 .87255 .87274 .91596
59 .05441 .05499 .05596 .06541
60 .01920 .01926 .06412 .06432
61 .53181 .54562 .54815 .58842
62 .04558 .04620 .04656 .07320
63 .33134 .33625 .33646 .37162
64 .30971 .31671 .32585 .35419
65 .35656 .35980 .36013 .39698
66 .03318 .03559 .04269 .09813
67 .16525 .19025 .19508 .28277
68 1.52184 1.54679 1.54771 1.60373
69 .20629 .20849 .20921 .22732
150
TABLE B-3--(Continued)
M W
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
Using Profit Method 5
1 .11741 .11818 .11965 .13456
2 .05324 .05370 .05495 .05813
3 .06031 .06152 .06793 .07000
4 .06925 .07114 .07242 .07260
5 .04003 .04015 .04186 .04279
6 .06371 .06530 .07804 .08016
7 .11635 .11909 .12092 .12593
8 .20787 .20843 .20869 .22326
9 .12816 .12943 .13152 .15636
10 .12186 .12323 .12449 .22120
11 .17104 .17190 .17278 .24052
12 .16467 .16477 .16541 .23462
13 .08378 .08390 .08530 .09710
14 .11694 .11733 .12008 .15850
15 .09549 .09595 .09700 .10402
16 .14974 .15102 .15150 .19027
17 .13802 .13834 .13918 .15022
18 .17273 .17675 .18090 .19518
19 .14597 .14687 .15194 .20336
20 .05987 .06197 .06406 .06975
21 .14248 .14286 .14371 .17643
22 .11266 .11282 .11327 .14434
23 .15842 .15880 .16235 .17352
24 .02336 .02339 .02416 .02950
25 .12663 .12708 .12809 .14690
26 .03125 .03142 .03540 .05245
27 .15938 .16012 .16376 .20174
28 .14847 .14903 .15154 .15931
29 .20847 .20947 .21020 .22687
30 .1824/ .18400 .18675 .22783
31 .10004 .10066 .10253 .13086
32 .07399 .07418 .07448 .07668
33 .04523 .04537 .04571 .04846
34 .11856 .11908 .11943 .12257
35 .07291 .07295 .07314 .07393
36 .05847 .05872 .05874 .05903
37 .15056 .15205 .15254 .19347
38 .13522 .13564 .13664 .16157
39 .16258 .16320 .16384 .19888
40 .11368 .11398 .11499 .13458
41 .13472 .13515 .13622 .16180
42 .13703 .13783 .13797 .18283
151
TABLE B-3--(Continued)
———'_'—‘—_"'M— W
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
43 .16897 .17013 .17027 .21333
44 .13230 .13248 .13259 .1840?
45 .12077 .12114 .12119 .18440
46 .10216 .10275 .10280 .17081
47 .10134 .10169 .10177 .13811
48 .19567 .19994 .20047 .22651
49 .08520 .08530 .08541 .10209
50 .14519 .14704 .14720 .17972
51 .11510 .11554 .11594 .14358
52 .09974 .10013 .10197 .10581
53 .16837 .16935 .16947 .17486
54 .09581 .09592 .09595 .09954
55 .08091 .08107 .18819 .08836
56 .06054 .06110 .06136 .08556
57 .12463 .12475 .12499 .12735
58 .61659 .63418 .63430 .66197
59 .04208 .04226 .04264 .04635
60 .02945 .02950 .05745 .05757
61 .35535 .36103 .36258 .38737
62 .03975 .03999 .04015 .05180
63 .26029 .26312 .26327 .28908
64 .25403 .25779 .26379 .28239
65 .21711 .21819 .21837 .23853
66 .05039 .05254 .05671 .08927
67 .13321 .14315 .14554 .18880
68 1.25848 1.27532 1.27607 1.32179
69 .14114 .14206 .14250 .15359
Using Profit Method 7
1 .14453 .14592 .14804 .16963
2 .07394 .07485 .07666 .08124
3 .06848 .07033 .07899 .08178
4 .08354 .08678 .08859 .08884
5 .04053 .04070 .04287 .04405
6 .07418 .07666 .09372 .09656
7 .09216 .09491 .09724 .10359
8 .31702 .31842 .31885 .34254
9 .14863 .15047 .15310 .18429
10 .14493 .14705 .14870 .27521
11 .20540 .20673 .20787 .29528
12 .23187 .23206 .23304 .33782
13 .09069 .09084 .09258 .10720
14 .13460 .13519 .13886 .18996
152
TABLE B-3--(Continued)
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
15 .09470 .09520 .09639 .10428
16 .16977 .17156 .17215 .21986
17 .16560 .16609 .16717 .18134
18 .19971 .20542 .21052 .22805
19 .18378 .18530 .19208 .26097
20 .06358 .06623 .06873 .07553
21 .17978 .18044 .18161 .22645
22 .12619 .12643 .12700 .16653
23 .17892 .17943 .18367 .19698
24 .01636 .01639 .01740 .02437
25 .14362 .14425 .14549 .16863
26 .02636 .02654 .03150 .05273
27 .19417 .19534 .20010 .24983
28 .17450 .17536 .17864 .18878
29 .22970 .23095 .23178 .25062
30 .22173 .22416 .22773 .28112
31 .11358 .11449 .11690 .15334
32 .10520 .10569 .10623 .11020
33 .06833 .06876 .06945 .07510
34 .16140 .16249 .16304 .16789
35 .11269 .11282 .11322 .11488
36 .08047 .08122 .08126 .08186
37 .16410 .16604 .16663 .21546
38 .17139 .17214 .17355 .20866
39 .20379 .20484 .20573 .25360
40 .13318 .13362 .13492 .16006
41 .16563 .16637 .16783 .20304
42 .16525 .16650 .16669 .22553
43 .24310 .24563 .24584 .31118
44 .14962 .14987 .15000 .21243
45 .13815 .13866 .13873 .21636
46 .11775 .11862 .11869 .20638
47 .13042 .13110 .13122 .18652
48 .23651 .24320 .24389 .27766
49 .09012 .09026 .09039 .11050
50 .17182 .17457 .17478 .21581
51 .13270 .13334 .13385 .16896
52 .09432 .09489 .09773 .10366
53 .09389 .09451 .09465 .10075
54 .06127 .06148 .06155 .07159
55 .08568 .08587 .09365 .09384
56 .12805 .13127 .13199 .19807
57 .12978 .12991 .13016 .13266
58 .86610 .90468 .90487 .94809
59 .06663 .06734 .06831 .07777
153
TABLE B-3--(Continued)
T r —= m
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
60 .03272 .03283 .07769 .07789
61 .55301 .56737 .56989 .61017
62 .05898 .05978 .06014 .08678
63 .34155 .34662 .34683 .38199
64 .36500 .37324 .38239 .41073
65 .36938 .37273 .37306 .40991
66 .06390 .06854 .07564 .13108
67 .21711 .24996 .25479 .34248
68 1.53106 1.55617 1.55708 1.61310
69 .20991 .21215 .21287 .23098
Using Profit Method 8
1 .14297 .14412 .14591 .16409
2 .26455 .27622 .28265 .29901
3 .07349 .07529 .08315 .08568
4 .07447 .07666 .07804 .07823
5 .04271 .04285 .04467 .04566‘
6 .08766 .09070 .10839 .11135
7 .11639 .11913 .12096 .12597
8 .23829 .23903 .23933 .25603
9 .12978 .13109 .13320 .15836
10 .12454 .12596 .12725 .22611
11 .17276 .17363 .17452 .24294
12 .16451 .16460 .16524 .23438
13 .08456 .08468 .08609 .09800
14 .12009 .12050 .12333 .16278
15 .11800 .11869 .11999 .12867
16 .12882 .12977 .13018 .16350
17 .13890 .13922 .14006 .15117
18 .17832 .18261 .18690 .20166
19 .14657 .14748 .15257 .20420
20 .06036 .06248 .06459 .07034
21 .14281 .14319 .14405 .17684
22 .11279 .11296 .11341 .14452
23 .15566 .15603 .15952 .17049
24 .02338 .02341 .02418 .02952
25 .13350 .13400 .13506 .15490
26 .03154 .03172 .03574 .05295
27 .16046 .16121 .16487 .20311
28 .14735 .14790 .15040 .15811
29 .20895 .20996 .21069 .22740
30 .18314 .18469 .18745 .22868
31 .10027 .10090 .10278 .13117
154
...... .- ,__.___....__.____._- ...—_H‘-... ___--_ .__—._._H ._ -——._...._..__..H__.._-_ -_... H...‘ - ...—1-. --- ._H___ -._._.-_ ..-..H.- .____. HH_. -..... _.__. HA. -.__.. ......
Adjusted for Adjusted for
Industry Unadjusted Adjusted for (I) and (I), (R) and
Number* (U) Intangibles(I) Royalties(R) Advertising(A)
32 .07576 .07595 .07626 .07851
33 .05836 .05859 .05903 .06258
34 .11953 .12006 .12042 .12359
35 .07353 .07357 .07376 .07456
36 .05889 .05915 .05917 .05946
37 .17698 .17904 .17962 .22782
38 .13908 .13953 .14056 .16620
39 .16363 .16426 .16491 .20018
40 .13114 .13153 .13270 .15531
41 .13679 .13724 .13832 .16430
42 .16806 .16925 .16943 .22452
43 .19400 .19553 .19569 .24519
44 .14952 .14975 .14987 .20800
45 .13567 .13613 .13619 .20722
46 .13903 .14011 .14019 .23293
47 .13349 .13409 .13420 .18211
48 .18796 .19190 .19241 .21740
49 .10105 .10120 .10133 .12111
50 .19413 .19745 .19766 .24133
51 .13810 .13874 .13922 .17241
52 .12473 .12534 .12765 .13245
53 .20075 .20214 .20229 .20872
54 .10412 .10425 .10428 .10819
55 .08921 .08941 .09725 .09744
56 .14056 .14361 .14423 .20111
57 .13857 .13872 .13898 .14161
58 .44413 .45317 .45326 .47303
59 .05609 .05640 .05691 .06186
60 .05336 .05355 .10427 .10449
61 .51490 .52691 .52917 .56534
62 .08277 .08380 .08413 .10856
63 .43796 .44603 .44630 .49004
64 .30709 .31262 .31989 .34244
65 .32509 .32751 .32779 .35805
66 .06624 .07001 .07556 .11894
67 .14702 .15923 .16188 .21001
68 1.44148 1.46362 1.46448 1.51695
69 .11556 .11618 .11654 .12561
*All business establishments refers to corporations (C), partner-
ships (P), and sole proprietorships (SP). The profit rates are based on
average returns for the five-year period, 1956-1957 to 1960-1961 for C,
and the four-year period 1957-1958 to 1960-1961 for P, SP. These profit
methods are outlined in Table 7 and Appendix A's section "Profit Rate and
'Welfare Loss'Adjustments” explains the distinction between Methods 1-8
and I-VIII. For Industry Coding see Table C—l.
SOURCE: See Table B-1.
APPENDIX C
ELASTICITY ESTIMATES FOR THE AMERICAN ECONOMY,
1956-1957 TO 1960-1961
155
156
ELASTICITY ESTIMATES FOR THE AMERICAN ECONOMY,
1956-1957 TO 1960-1961
Dorfman-Steiner-Telser Propositionl
This proposition states that a profit maximizing firm selects
a price and advertising budget such that the price elasticity of
demand equals the value of the marginal sales effect of advertising
(v.m.s.e.a.).
Assuming, as usual, continuous and differentiable functions:
Let (1) C = g(q) represent total production cost (C) as a function
of the rate of output (q) i.e., the quantity the
firm can sell per unit of time
2
||
(2) i(a) represent the number of consumers made aware of
the product (N) as a function of the firm's
advertising budget (a). Since the cost of making
N consumers aware of the product is a, the mar-
ginal cost of awareness (m.c.a.),we get by
implicit differentiation
(3) m.c.a. == 1/(di/da) Since sales (q) of the firm may vary
directly with N and inversely with price, p, the
demand function is
(4) q = f(p,N) The necessary conditions for maximizing net
revenue, (R.= pq-C-a), means that the firm must
select its p and a so that
"(5) dR/da = p [df/dN di/da] - dg/dq df/dN di/da - 1 = 0:
(p - dg/dq) df/dN di/da - 1
(6) dR/dp = P df/dP + <1 - dg/dq df/dp= 0= (P - dg/dq) df/dp+ q
The elasticity of demand (It) is
(7) 371. = - p/q df/dp = p/ (p - ds/dq)
The v;m.s.e.a. (where m.s.e.a. = p df/dN and
df/dN is probably not constant since not all
consumers made aware purchase the same amount--
if at all).
1
R. Dorfman and P. O. Steiner, "Optimal Advertising and Optimal
Quality," American Economic Review (Dec., 1954), pp. 826-836; Lester G.
Telser, "How Much Does It Pay Whom to Advertise," American Economic
Review, (May, 1961), pp. 194-205.
157
(8) v.m.s.e.a. = p df/dN di/da = p/ (p - dg/dq)
Thus’n’ = v.m.s.e.a. If we define the ratio
of 1 - marginal production costs (m.p.c.)
to the price to be the marginal advertising
intensity (m.a.i.), we get
(9) m.a.i. = (p - dg/dq) / p The rule for optimal advertising
expenditure is at the intersection point of the
marginal return from awareness curvel:m.r.a.=
(m.a.i.) (m.s.e.a.)] and the marginal cost from
awareness curve (m.c.a.)--the former curve
probably decreases as N increases while the
latter increases.
(10) (m.a.i.)(m.s.e.a.)== m.c.a.
If we assume m.p.c. is approximately constant
[==average variable cost (a.v.c.)], m.a.c.fitfl.-
a.v.c./p. Also note that total revenue is equal
to total variable costi'a + R. This means the
advertising intensity can not be greater than the
m.a.i.
(ll) m.a.i.%(p-a.v.c.)q = R+a 2 a/(pq), R20
Pq Pq
Since from (7) and (8) we know that m.a.i. and
'flvary inversely, ncan not exceed the recipro-
cal of the advertising intensity as taking re-
ciprocals reverses direction of the inequality.
(12) 72 = 1/m.a.i. 5 1/(a/pq)
In economic terms, this means that if a.v.c. is
constant, then the reciprocal of the advertising
intensity is an upper bound to the price elasticity,
e.g., if advertising expenditures are fifty per
cent and one per cent of total sales, i.e., ad-
vertising intensities of fifty and one per cent,
price elasticity at the optimal output is be—
tween one and two in the first case and less than
100 in the second. Although the proposition
strictly applies only to differences between
upper bounds of the price elasticities, it is not
unreasonable to expect that there may be similar
dispersions among the actual elasticities. We
can also formulate the proposition for the case
where m. p. c. > a.v. c.
158
(13) 1/7b= m.a.i. = 2 - dgzdg< (p-a.V-c.)/p = (HM/(m)
p
(14) 1/7Z<(p-a.v.c.)/p>8/(pq) If we take reciprocals we
must again reverse the direction of the inequality.
(15) 7?. > lamp-a.v.c.) < (pq)/a
Thus,because the average advertising intensity is
probably closer to the m.a.i. under increasing a.v.c.,
the ratio of sales to advertising outlay may be
even closer to at for increasing m.p.c. (although no
longer an gpper bound).
Again, in economic terms, the above analysis means we should
expect to find lower price elasticities with heavily advertised
products, such as drugs, tObacco, beverages and toilet preparations
than with little advertised products such as lumber and textiles.
Ideally, of course, we should like to have some independent estimates
of product elasticities to see if our proposition yields useful
results. Unfortunately, such information is in the nature of a
scarce good.
There are a couple of things to be noted about our table of
elasticities. First of all, we used the more easily obtainable
figure for business receipts (gross sales and gross receipts from
operations) as contained in the Statistics of Income in place of
sales-~the difference is typically quite slight. Secondly, we have
computed the elasticities only for corporations and not for sole
proprietorships and partnerships, since the above source does not
contain this information. Thirdly, we have computed the elasticities
on a three-year basis, July,l958-June,l959, July, 1959-June, 1960 and
July, 1960-June, 1961 (as well as on a yearly basis), in order to
159
avoid, as much as possible, ephemeral factors while remaining within
the substantially revised Standard Industrial Classification employed
by the above source. Finally, we want to reiterate our statement, so
as to avoid any misunderstanding, that our calculations involve
maximum or gpper bound elasticities and not the actual elasticities
themselves which might be anywhere between one--the lower limit by
classical arguments--and our gppgr bound estimates. Since advertising
costs are typically well under two per cent of sales for most indus-
tries, we should not be surprised at what appears to be rather large
numbers for our maximum elasticity estimates (a two per cent ratio
is associated with a '7ziof 50). What is important is not the
absolute estimates but the relative dispersions among industries.
Perhaps some sort of coding on the basis of known elasticities might
make them useful for absolute purposes too. The elasticity estimates
utilizing this approach are shown in Table C-l.
160
m.N¢H H.NmH o.omH N.QMH moHuumomcH moHHHm
ecu chwanHnoe chHuoHum .mH
m.sHH a.mHH N.mHH m.NHH muuomoum moHHHm vow modem .NH
m.aN m.Hm o.oN a.mN monoume use ououchnm .oH
m.HMN m.HmN H.m¢N H.N¢N ououchsu udooxo
«muoomoum @003 mam umnEoH .mH
N.eoH N.NoH N.moH a.moH mHmHuouma umHHEHm mum
mowunmm Eoum mums uuoomoua
moanch nonuo mum Honmmm< .eH
a.msH m.csH w.esH a.meH nauseous HHHa «HHuxoe .mH
m.wH a.mH a.mH a.mH muousuommscms ouomDOH .NH
N.a¢ o.Nm o.Nm n.Nm muuomoum moumcHx mom moon .HH
a.mH m.ca o.oN o.o~ mmHHHmaecH smssm>om .oH
a.mo m.HN ¢.HN o.HN wuHuouomwscmE Houom. .m
¢.0m¢ m.mNe H.mm¢ 0.0mm coHuosuumooo .w
H.ms~ ”.HoN ~.mNN N.¢mN wchHs mHHosHaucm
mom «mHmuocHE oHHHmuoaucoa
mo wcwmuumov use onon .N
a.mmm N.om¢ N.coo m.¢m¢ new
Hmuoumc mum EsoHouuom omouo .o
a.msH.H m.aao.H H.s~o.H o.Nmo.H mchHa
ouHcmHH mom Hmoo moosHasuHm .m
~.-~.~ m.ass.m a.msH.N m.mmm.m wchHa Hans: .s
m.Ne¢ m.cqm N.ooo m.on wchHE Hmuoa .m
m.OON m.oNH H.ooH o.wNH moHuoanm
mom Nuuwouow «ououHooHuw< .N
4.9m a.mm m.am N.Nm masons HmHuumsecH HH< .H
HomHuoomH oomHummoH mmaHuwmmH mmoHumm mouaH
muHoHummHm u MNN huHoHummHm u NNN. huHuHummHm u Hyw ecu HON muumsmcH
NHHoHumem
owmum>¢ n mwfi
In
kmuQ< MmmHmyammzHMHmuz<2hmoa
mmh UZHmD mmH¢ZHHmm nz
sauce use eoHoHne> some: .mN
N.HNm a.mmm o.wm¢ o.¢wm moHano> uouoa uaooxo
«usoaaano ooHumuuommaeuH .QN
a.mm m.Nn ¢.om H.mn moHHemse one uses
. aanao «muocHnomB Honuuoon .NN
o.Nm H.mm a.mw N.¢m uooaaH=ae soHueuuoauseuu vow
HmonuooHo annexe «muooHsomz .oN
o.~cH m.HHH «.NQH a.moH uaoaaHssu aoHuosuoa
names» was huoanusa.umooXo
«Housmnmuo wsHmoHuoHv
nauseoua Heuoa moumuHunmm .mN
N.NNN a.mNN o.NMN a.mNN eoHuuuamsH Hmuoa huwaHum .eN
N.mnH «.NeH n.nnH «.NmH nauseous
umme use ahmHu «unaum .NN
w.Nm N.mm e.oa N.am nauseoua uonueoH use uunumoa .NN
a.mm m.co a.mn N.Nm nauseoum eOHueeHd
mooosoHHuoeHa one means: .HN
o.an n.wHN N.N¢N H.¢oN uoHuuunmuH voumHou
mom wdHonou aboHouuom .ON
m.m~ a.o~ m.aw o.m~ nauseous
moHHHe use «HuoHaunu .aH
HomHaoomH n comHammmH N mmmHaommH H evoHuom sonny
NuHuHuaaHm u ~N. NuHuHHmuHm.u NN NHHuHuasHm u en N on» How NuumaeaH
UHoHuuon
owmuu>< u QNA.
:1
HeosaHuaoov--H.o mamas
162
N.oo H.HN H.mN o.HN oumumo Hmou mum
HoocmuomoH «moomch Houoa .Nm
H.ooH n.HoH m.am ¢.ooH oHomooHHm uoc
ommuu HHmuou use onmoHonz .Hm
m.HN m.mN a.mN m.sN monOHm HHmoou nonuo .om
m.NNH m.NeH N.wNH m.NNH ucoEmanm Baum use
masseuse .mHmHuoumB wuHmHHsm .me
m.aoH a.moH a.mHH ~.NoH mausHa waerHue ens maHHsm .me
m.oHH N.cHH ¢.MHH ¢.MHH acoHumum m6H>umu ocHHommw
mom euonom o>HuOBOu=< Ne
3 . 4 o m on N . on w . om ucmEn—anm van
mwcHsmHsuom mac: Hououwouoh .oe
m.Hs a.me N.Hs c.~s maHHommouus ecu Hosanna .me
a.mm m.o¢ m.am m.am omHmamnouoa Hmuucoo .ee
n.HoH m.aoH N.NHH a.moH mock .me
N.No a.mo a.No o.wo ommuu HHmqu .Ne
a.moH N.mmH a.moH N.NoH uuuHmmuHona nosuo .Hq
H.mnH 4.434 n.\NH ¢.n¢H mmHHdoum mom unoEmano
wcHumos mam mcHoenHa mom
«oum3mum: ammoow HmoHuuoon .oq
N.mwN N.HmN a.mHN a.meN muosmoua moumHun mam uoHuooouu .mm
o.mNH a.moH a.mNH H.NNH ommuu onmmHo:3 Hmuoa .wm
m.Nm o.mm a.mm m.am ommuu
HHmuou mom onmoHosa Hmuoa .Nm
H.oHo.H m.mmm ~.soN n.0sm mmuH>usm NHmHHcsm
nozuo com xHemom nouns .om
a.mms a.moe a.H~e N.m~s «saunas ecu
moHunmaoo new use UHHuoon .mm
m.cmH o.HmH m.awH c.NmH coHumoHcoeaoo .em
m.¢ON m.NON a.mON a.mON coHumuuommcmuH .mm
meHaoomH oomHnmmmH N mmmHuwmmH H mmoHuom sunny
NHHoHHuaHm u mNN NHHuHumaHm u NA. NHHuHumsHm u NNV «as sou NuumsecH
NHHuHuaaHm
oweuu>¢ u mwfi.
l n g
HeuseHueoov--H-o mumae
163
m.aN H.Hm a.mN H.wN mmHsooHa :oHHoz .oe
w.Nm m.coH m.aoH m.ooH mooH>uom uHmdou
nonuo mom amowmumw mum
mosH>uom «Manama mHHnoEOusd .mo
m.wNH N.moH N.Nm a.moH mooH>uom mmocham .qo
m.co m.am ¢.oo H.oc mmoH>uom HmGOmuom .mo
H.He m.an m.ce ~.os amusHa
wcmeoH nonuo mom «enema
«amazon wcHaoou amHouoz .No
N.mm a.mm N.mm a.me mooH>uom kuoH .Ho
H.mmN a.moH N.NoN m.amH mwuHmHHoo umooxm
azuuomoum Hmou mo muOmmoH .oo
N.NN «.NN m.cm ¢.oN mwchHHon
cmsu umsuo zuuoaoua Hmmu mo
muoumoH udmuXo aoumumm Hmom .mm
q.ow a.mN a.mo e.¢N ooH>umm mam
«muoxoun «mucowm moomuomcH .wm
c.Nme o.oom N.mmm a.mNm mumHuumo oocmuomuH .Nm
«.NH a.mH N.mm w.wH mooH>uom mum «newsmnuxo
«muonom amuoxoun
%UHmOEEoo mam mUHuooom .om
¢.oo N.mm e.om ¢.aq moHcmdBou uses
numo>cH nonuo mom wchHom .mm
N.mH H.0H N.wH 0.0H exude cosy
nouuo moHuomwm uHmouu .em
o.m o.m m.m N.m wcermm .mm
HomHuoomH m oomHummmH N mmmHuman , mmoHuom conga
muHuHuumHm n N muHoHummHm .... N huHoHummHm u me 9.3 you muumomcH
NHHoHHmmHm
owmuo>< u «NV
Ammmmwmmmmv--H-o mHmaa
164
.wummm unm>oHou can now «ooH>uom mouo>mm
HmcuouuH «moosuumamo zuommoua .m .D .msuouom me uaoocH coHumuomh w .oEoocH mo moHumHumum HNUMDOm
A mmoHuoa mounu use How mmmuo>m u m~fi
A
A $2 83.182 .HHE. u in
.AHco some coHumuoauoo A
no momma mum moumEHumo madam. A 003 mcohsammmH .33. u NNA
A .
A mmmH sage--mHAH aHsa u Han
x
No.4: u m> $3.2m n ~> mNo.5” u H> Us . Le .... > 83329 .8 253328 9:.
m.No N.wm o.Ha m.mN oHomooHHm
uoc mmochon mo muoumz .mo
v.mo N.mo o.Hm m.Ho mooH>uom nonuo .wo
o.wm m.wm o.mm m.mm monouoHa
coHuoE umooxo «mooH>uom
coHummuoou mum ucosomsad .Nc
HmmHnoomH oomHnmmmH mmmH1wmmH mmoHumm omega
33335 n m N ESTES u «N 33335 n Hg .9: new A
HumomcH
NHHoHHmmHm
owmuo>< u mmfi.
AmoouHucoovuaHuo me HH> H0. 0 >H HHH HH H
0000002 ammomm. tympanz
mm .m 00 oaoouH vexmuus mm .m 00 esoooH vexmuc: hummuuuH
nqu 0 mo maoooH me COHumHOQHOUumuowmm
zuHs 0 mo oaoocH me GOHumHOQHOUuuoum<
emo¢cmmm< ZOmzH0001MMZMmH NEH wszD mMH¢SHHmM 020200 mo MHHUH80¢HM MUHmm
N10 mHm¢H
169
00.0 0N.0 0N.N 0H.0H 00.0 NN.N 00.N 00.0 00
00.NH 00.0H 0H.0N 00.0 0N.0H NH.0N 0H.N0 0N.0 N0
0H. 0H. NH. 0H. NH. 0N. HN. 0H. 00
N0.N 00.H 00.H 00. 00.0 00.H 00.N 00.H 00
0N.0 NN.0 H0.0 00.0 0N.0 00.0 NH.0 00.0 00
0N.0N 0N.00N 0N.NOH N0.HOH 00.HNH 00.00 00.00H 00.N0 N0
00.00 0H.0HN 00.0 00.0 0H.00 00.0N H0.0H N0.0 00
NH. 0H. NH. 0H. 00. 0N. 0N. 00. 00
00.N NN.H 00.H 00.N N0.N 00.0 0N.0 00.0H 00
0H.H 00.H 0N.H N0.H 00.H 00. N0.H N0. 00
HH.00N 00.N00 00.0NN.0 00.NON.H H0.0N0 00.0N0 0N.0NN N0.00N H0
00.0HH 0N.N0 0N.N0 N0.00 NH.00 NH.H0 0N.00 0N.H0 00
N0.N0 NN.00 00.00 N0.00 00.00H H0.00H 0N.00N N0.00H 00
00.00 H0.00 00.0N 00.00 0N.0N 00.00 0N.NN 00.NN 00
00.000 NN.000 NN.H00 00.00N N0.000 0N.N00 00.00N 00.00HHH N0
00.00 0N.0HN 00.0N0 0N.00H NH.0H0.H NN.00H 00.00H N0.000 00
0H.000 00.000 0N.000.NN 00.HON 00.NON NN.000 0N.NNN 00.00H 00
00.000 NN.000 00.N00 00.00H 00.00H 0H.00H 0N.NHH 0H.NNO 00
00.0N 00.0HH 00.00 0N.00H 00.0N 0N.00H N0.00 NN.00 00
00.00N N0.000 0N.000 00.H0 N0.00N 0N.0NN.N 00.0N0 0N.HON H0
0N.0H0 0N.000 0N.00 0N.N00 00.0NN 00.0NN 00.00 0H.0H0 00
H0.00N 00.000 H0.000 N0.NOH HN.HON 00.00N N0.0N0 00.00N 00
00.N 00.N 00.N 0H. 00.0 00.0 0N.0 00.0 00
00.NN 00.0 0N.0 00.N N0.0H 0N.0 0N.0 00.N 00
00.00 00.0N 00.NH 00.000 00.00 00.0H N0.HH 00.0N 00
00.0H 00.0 0N.0 0N.0 00.0H N0.HH 00.0H 00.0H 00
0N.N0 00.N0 00.00 0N.0NH 00.00 00.00 00.00 00.HN H0
HHH> HH> .H> > >H. HHH HH H
avocuoz uHmoum eumneuz
muumuvcH
nqu 0 Ho mEoocH me :OHumuomuoonmuomom
mm .0 mo meoocH woxmucs
mm .0 mo meoocH vmxmucm
:qu 0 mo mEoocH me :oHumHOQHOUIHmu0<
HamscHseoov--~-u HHHHH
170
.Ham mHan mom "000000
.Huo oHan mum 0cHwoo umpEoz muumsvoH Homes
.Hm .m use HoaH-oomH on wmmH-NmmH .eoHHma ssmN-HsoH was ens .o How .HomH-OOAH on NmmH-0mmH .eoHHsa
ummhum>Hw on» now mauSuou mwmuo>m commonumss :00: ummmnqummv mmHsmHOHmHuaoum mHOm was .Amv maHnmuosuumm
.on mCOHumuoauooH muauaanHnmumo mmoonsn HHm you AN mHan ommv HHH>1H mvonumz unoua mch0¥
N0.asm n m> N~.osm .u s> em.ssH u 0> NH.NmH u N>
No.0me u N> No.0mm n m> N0.Hmu u m> Na.msH n H>
x
"mum .ImHWI n"> .:0HumHum> mo mucoHonwmoo mLH
00.0N 00.00 0H.NON 0N.00 0N.0H 00.0NN 0H.00 00.0N 00
00.N N0.N 00.N H0.N 00.N 00.N N0.N 00.N 00
N0.00 00.NHO 00.00 N0.00 N0.N0 NH.NON 00.N0 00.00 N0
00.0H 00.0H 0N.0H HH.0H 0N.0N 00.NH H0.0H 0H.0H 00
0H.NH 0N.NH 0H.HH 00.0H H0.0H N0.0H 0H.0H 00.NH 00
N0.0H 0H.0H 00.0H 00.0H 0N.0H 00.0H H0.0H 0N.0H 00
HHH> HH> H> > >H HHH HH H
muonumz unoum «aumpeuz
mm .0 mo esooaH vexmuca mm .m mo mEOUCH vmxmucs hummunuH
nuHa 0 mo maoosH me soHumuoauoormuowom
nuH3 0 mo maoocH me GOHumuoauoonuou0<
HemacHHaoov--N-u mamas
000000.H
0000N0. 000000.H
N00000. 0NO0H0. 000000.H
000N00.1 000000.: 00NH00.1 000000.H
NON000.u 0H0000.1 HNN000.1 00NNH0.1 000000.H
0N0000.n NHN000.1 00N000.u 0NN000. 0HO0H0. 000000.H
N00000.n 0HOH00.a 0N0000.1 000N00. 00H000.1 00NON0. 000000.H
0N000H.u 00HN00.1 H000HH.- 00NON0. 000000. HNOHOH. 0HNO0H. 000000.H
0NO0N0.u 0N0000.1 N0HNN0.1 000000.1 0NO0H0. HH000H. 0NO0H0. 0N000N. 000000.H
0H0000.1 0HH000.1 0H0000.n N00000. 0NO0N0. 00000H. NONN00. NONHON. 000000. 000000.H
00NNNH.1 000N0H.1 0000NH.1 N00000. HNN000. 000NNO. 0000NN. 000000. 0N000N. 000NON. 000000.H
HHHVx HoHvx Hmvx Hmvx Hvi Hove Amvx Have Hmvx HNVx AHvx
HHHvx
HoHvx
Hmvx
Hmvx
HNVx
Hove
Hmvx
Have
Hmvx
Hvi
HHVx
mHanum>
xHuumz ucoHonwooo coHumHouuoo ucoaoz moonoum novuo ouoN
00uz
H00H1000H anomouam< 0uHmcmucH 0cHwHuuu>v< nomHoHuuoaHoumucmEmuoo com: 0omm0 mumEHumm muHoHummHm
000H1000H acomouam¢ AuHmcoucH 0cHwHuuo>0< nowHoHauocHoumucmbuoa com: uowmm oumEHuwm muHoHummHm
000H1000H aHomouanz huHmcoucH 0onHuuo>0< ammHeHuuocHoumncmawuoo com: vommm onEHumm muHoHummHm
HHH> 000mm: unotm coo: cmumm mumEHuuu auHoHummHm tcacmncmutmcuwH
HH> vosuoz uHmonm com: vmmmm oumaHumm %UHoHummHm COmCHQOMauocuoH
H> conuoz unoum comb voumm oumEHumm muHoHummHm comcHnoMnumcuoH
> voSumz uHmoum com: 0mmm0 oumEHuwm huHoHummHm concH0001uoouoH
>H vogue: uHmoum com: 0omm0 oumEHumm huHoHummHm momcHnoMIHocuoH
HHH wonumz uHmoum com: 0omm0 mumEHumm muHoHummHm comchomnuocuoH
HH vonumz uHmoum com: 0mmm0 oumEHumm muHoHummHm somcHnomuuocumH
H vogue: unoum com: 0ommm oumeHumm muHoHummHm somcHQOMnumcuoH
171
HHHVx
HoHvx
-Hmvx
Have
HNVx
Hove
Hmvx
Have
Anvx
Hmvx
HHvx
mmHanum>
E "I
mm,k<2.kmm .HHHUHHmch «Ho ZOHHHNHMMHHOU
010 MHm¢H
172
000000.H
000N00.
0000N0.
H00000.1
N0000H.u
0N0000.u
000000.
H00000.1
0000NH.1
NN0N00.1
00N0N011
HHHvx
HHHVx
oooooo.H . HoHvx
Newcom. oooooo.H Hmvx
@0Hmoo.- NNNmoo.- oooooo.H vax
smNNso.- mommoH.- mmmemm. oooooo.H Hvi
mmNNmo.- mHSHoH.- Nmmsmm. omowom. oooooo.H Hove
snowmo. Hmsomo. smmmoa. Nommom. «memos. oooooo.H Have
Nosmso.- NsmHmH.- wNHosm. oNHNmm. mmmmmm. mammsN. oooooo.H Hsvx
Nstmo.- «mumsH.- mNmaeN. mmmmwN. eNHnsm. umquN. nqusm. oooooo.H Hmvx
monoo. sHmmNo.- mesz. somHmm. «Samoa. mNmmoN. staom. smmsmm. oooooo.H Hmvx
m0m~m0.. NmmNHH.- ammomN. mosHNN. Hosssm. oNNmNN. awmaHm. Hmommm. mHosem. oooooo.H HHvx
HoHvx Hovx Hmvx Hvi Hove Hmvx Have Hmvx vax HHvx mHHmHHs>
xHuumz uanonmmoo GOHumHouuou xcmm
I {H
AvonoHuooovau0au 0H00< 0cm Amv mmHUHmhom AHv mmeHmcmusH momma
AeV.HHV How emumsHe< 0:0 HHV How 0000000< H00 00000000 HHmosm 0000000000
AwumHHov commoonuv
mHZMZHmDHQ¢ madHQMZMMHzH HH¢ quQDHUZH >H nomHmz HHmomm mom :mmmmOH mm¢mqm3: WMHmDQzH10010080302H
Hun MHm0< 000 va 00H0H0>oH AHV 00H0000000H 00000
va .000 you 0000:00< 0:0 AHV you 0000:00< 000 00005000 uHmoum 0000:0000:
H0000H00000--H-0 04000
177
can uumm unoum: newuoom m.< chammm< cum 0 wanes mum
.>H UOSUOZ UHNOHQ MO GOHUQHHUmUU m HOM.MquEumDHU< .mMOA UHMMHQS.
.m mfinmfi mam “038:2 muumsvcH we wcfivou Hom«
000.N00.0~ 000.H00.0 000.000.0H NH0.0H0.0 0~0.000.0H 000.000.0 000.000.0H 000.000.0 000MMMMH
Z
HmuOH
H00.00 000.~ 000.00 H00.H 000.00 000.H N0~.00 NNN.H 00
0HH.000.0 00~.000.H 0H0.0H0.0 00~.000.H H00.0H0.0 000.000.H N0N.~00.0 00~.000.H 00
0N0.00 00H.0 HOH.~0 0N~.H 0NH.H0 0H~.H 0~H.0~ N00 00
~HN.0N 0~0.0 00~.~0 000.~ 0N0.00 00N.~ 000.00 0NH.0 N0
~00.~00 0H0.00 0N0.H0~ 00~.N~ 000.00~ 00N.0~ 0H0.N0~ 000.0~ 00
...— " 1H l .H .l .H H
H H u H I H «000502
NN, NR ND. MW NR. NR. NR. huumsncH
N00 00H0H000>00 000
N00 .NH0 000 00000000
N00 00H0H00o0
vow AHV How vmumsmv<
AHV mmanflwcmucH mmumm
How vmumsmv< uwwoum vmumsnvaD
~0000H00000--H-0 04000
178
TABLE D- 2
INDUSTRY-BY-INDUSTRY "WELFARE LOSSES" FOR PROFIT METHODS
II, III, V-VIII, UNADJUSTED AND FULLY ADJUSTED RESULTS
(thousand dollars)
Industry Unadjusted Fully Adjusted
Number* (7?: 1) (7(1) (7?: 1) (771.)
Method II
1 129,354 1,456,258 161,625 1,627,805
2 18,149 139,389 17,533 137,003
3 6,841 90,585 7,259 93,312
4 30,080 297,888 31,813 306,348
5 621 25,310 1,083 33,400
6 33,179 910,742 43,210 1,039,343
7 2,021 90,617 21,423 295,029
8 2,956 267,495 23,188 749,249
9 289 25,658 7,082 126,956
10 10,430 265,420 15,092 319,276
11 322 44,424 1,868 107,047
12 489 45,754 1,101 68,667
13 360 30,186 1,697 65,526
14 4,820 163,747 8,526 217,780
15 10 7,881 382 49,785
16 2,652 180,207 37,621 678,991
17 82,157 1,220,445 95,927 1,318,764
18 1,423 69,694 4,886 129,142
19 685 36,022 1,827 58,804
20 2,327 106,479 4,613 149,934
21 18,922 505,521 24,781 578,519
22 3.845 189,079 8,909 287,793
23 6,822 299,531 15,839 456,409
24 2,563 161,800 11,991 350,000
25 5,102 201,627 6,521 227,942
26 243 53,816 2,397 169,174
27 421 34,603 3,856 104,726
28 2,530 98,911 5,869 150,651
29 144,796 1,524,255 166,553 1,634,764
30 48,786 562,034 59,078 618,480
31 364,026 1,729,020 391,970 1,794,157
32 17,880 58,242 18,862 59,821
33 61 31,843 571 97,544
34 2,031 120,929 3,718 163,631
352 65 57,377 3,384 412,538
36 5,561 372,436 13,813 586,969
37 1,036 116,197 13,411 418,150
38 114 25,895 4,985 171,274
179
TABLE D-2--(Continued)
Industry Unadjusted Fully Adjusted
Number* 0z= 1) < 72 L) Oz= 1) (74 L)
Method 11 (Continued)
39 392 43,135 8,627 202,308
40 579 121,360 5,101 360,087
41 13,771 313,675 21,232 389,497
42 88 23,641 943 77,559
43 8,164 324,816 20,287 512,035
44 130 33,196 1,775 122,756
45 642,659 688,751 859,185 796,371
46 104,856 339,614 172,936 436,145
47 1,190,892 270,567 1,411,536 294,567
48 4,059 59,321 30,128 161,608
49 423 65,122 1,115 105,716
50 98,216 409,501 111,778 436,860
51 483,323 1,153,871 652,382 1,340,570
52 467,648 100,394 491,472 102,920
53 1,919 61,745 6,428 113,023
54 69,046 509,301 89,896 581,067
55 17,585 290,324 23,885 338,358
56 29,925 305,066 39,414 350,109
57 4,919 80,246 12,301 126,895
58 146 14,276 2,543 59,607
59 1,448,148 3,511,246 1,567,867 3,653,501
60 533 18,236 1,037 25,420
Total
”Welfare
Loss" 5,541,386 20,354,723 6,776,113 26,441,727
Method III
1 86,494 1,190,809 113,528 1,364,269
2 17,197 135,658 10,132 104,150
3 5,988 84,750 6,370 87,417
4 20,968 248,709 23,009 260,532
5 497 22,652 915 30,728
6 21,892 739,784 30,341 870,921
7 1,501 78,089 1,767 84,731
8 2,175 229,450 20,860 710,645
9 111 15,917 6,021 117,062
10 8,523 239,939 12,724 293,159
11 273 40,906 1,743 103,417
12 429 42,830 1,001 65,456
13 492 35,279 1,971 70,622
14 3,258 134,617 6,361 188,117
15 3 4,474 581 61,404
180
TABLE D-2--(Continued)
Fully Adjasted
Industry Unadjusted
Number* (7L=1) ( 7L1.) (71:1) (711.)
Method III (Continued)
16 1,107 116,493 29,053 596,682
17 68,171 1,111,719 81,026 1,212,016
18 985 57,976 4,030 117,282
19 525 31,521 1,554 54,237
20 1,445 83,919 3,313 127,057
21 13,931 433,757 18,932 505,662
22 2,750 159,912 7,175 258,282
23 4,667 247,753 12,418 404,115
24 1,629 128,990 9,822 316,765
25 5,985 218,363 7,491 244,311
26 13 12,400 1,357 127,272
27 194 23,506 3,367 97,867
28 1,965 87,174 4,871 137,247
29 120,993 1,393,345 140,636 1,502,198
30 36,092 483,412 44,855 538,912
31 295,109 1,556,772 319,225 1,619,131
32 14,633 52,689 15,488 54,207
33 30 22,207 601 91,331
34 148 32,604 763 74,135
35 10 22,149 2,819 376,571
36 2,482 248,827 8,672 465,093
37 484 79,423 11,124 380,829
38 38 14,924 4,373 160,404
39 194 30,368 7,583 189,668
40 165 64,724 3,641 304,219
41 5,237 193,435 10,264 270,811
42 330 45,872 1,570 100,066
43 6,640 292,929 17,842 480,188
44 78 25,734 1,566 115,326
45 808,046 772,307 1,044,494 878,061
46 58,622 253,933 110,705 348,957
47 753,418 215,207 883,610 233,061
48 1,569 36,885 22,326 139,118
49 1,144 107,086 2,150 146,820
50 84,675 380,225 97,445 407,890
51 1,091,478 1,733,988 1,346,716 1,926,090
52 395,803 92,361 416,759 94,774
53 2,912 76,070 8,150 127,264
54 63,065 486,741 83,079 558,663
55 16,883 284,4/5 23,047 332,369
56 26,100 284,906 35,061 330,213
57 4,489 76,659 12,797 129,427
58 26 6,017 1,864 51,033
59 1,389,295 3,439,156 1,507,344 3,582,290
60 12 2,770 166 10,178
Total "Welfare
Loss” 5,453,365 18,765,575 6,608,370 24,630,725
181
TABLE D-2--(Continued)
Industry Unadjusted Fully Adjusted
Number* (}z=1) (71L) (71:1) (K L)
Method V
1 385,872 2,515,184 442,030 2,691,993
2 7,803 91,400 8,517 95,488
3 6,082 85,412 6,442 87,908
4 17,763 228,915 20,325 244,868
5 1 1,059 97 10,015
6 21,055 725,510 29,322 856,170
7 21 9,306 11,383 215,056
8 5,720 372,119 30,121 853,952
9 1.992 67,337 12,411 168,065
10 2,622 133,079 5,109 185,766
11 0 732 13 8,889
12 767 57,288 1,505 80,269
13 310 27,996 1,588 63,397
14 1,325 85,855 3,470 138,938
15 4,656 173,827 8,191 230,547
16 5,389 256,978 44,254 736,415
17 47,623 929,188 59,901 1,042,108
18 544 43,075 3,059 102,182
19 6 3,288 354 25,880
20 3,755 135,263 6,481 177,716
21 66,690 949,054 76,534 1,016,691
22 209 44,107 2,191 142,725
23 38,352 710,200 56,149 859,328
24 3,837 197,980 14,565 385,738
25 828 81,198 1,421 106,389
26 25,780 554,760 37,395 668,143
27 3,918 105,566 10,842 175,608
28 233 30,032 1,743 82,094
29 235,651 1,944,524 263,352 2,055,641
30 25 12,832 749 69,657
31 146,072 1,095,259 161,324 1,151,023
32 5,192,365 992,518 5,516,245 1,023,005
33 425 84,203 1,401 152,819
34 7 7,292 332 48,900
35 5,961 547,555 39,452 1,408,651
36 2,457 247,567 8,621 463,723
37 550 84,643 11,416 385,789
38 10 7,531 3,970 152,838
39 257 34,936 7,958 194,299
40 405 101,485 4,735 346,901
41 5,848 204,413 11,070 281,239
42 1,437 95,712 3,516 149,730
43 1,730 149,534 8,846 338,111
44 6 7,034 1,099 96,620
45 706,304 722,051 928,917 828,057
46 136,205 387,066 208,560 478,965
182
TABLE D-2--(Continued)
Industry Unadjusted Fully Adjusted
Number* (71 =1) ( 7( L) (7(=1) (71 L)
Method V (Continued)
47 1,942,937 345,595 2,129,484 361,806
48 58,926 226,013 123,502 327,203
49 498 70,694 1,191 109,263
50 106,747 426,915 120,678 453,919
51 3,072,408 2,909,230 3,495,056 3,102,885
52 1,971,431 206,129 2,037,544 209,557
53 49,985 315,167 67,854 367,205
54 36,546 370,533 52,413 443,738
55 19,569 306,266 26,205 354,410
56 13,834 207,420 20,641 253,367
57 7,616 99,847 16,608 147,445
58 149 14,414 2,561 59,814
59 1,355,197 3,396,690 1,471,865 3,539,881
60 257 12,663 624 19,723
Total "Welfare
Loss" 15,724,967 24,249,442 17,643,204 30,828,523
Method VI
1 103,990 1,305,699 135,253 1,489,094
2 14,770 125,747 15,630 129,355
3 10,337 111,355 10,848 114,073
4 18,815 235,593 21,472 251,677
5 872 30,005 1,452 38,717
6 31,698 890,186 41,579 1,019,535
7 720 54,090 907 60,690
8 993 155,051 16,955 640,683
9 142 17,992 6,259. 119,351
10 10,186 262,301 14,817 316,356
11 251 39,281 1,700 102,126
12 749 56,629 1,492 79,915
13 23 7,587 618 , 39,537
14 541 54,872 2,165 109,736
15 1,289 91,457 3,545 151,669
16 211 50,874 23,393 535,413
17 118,445 1,465,396 135,317 1,566,293
18 71 15,573 1,669 75,474
19 378 26,746 1,263 48,904
20 35 13,143 675 57,351
21 125,012 1,299,379 139,162 1,370,949
22 707 81,079 3,533 181,230
23 72,664 977,571 96,832 1,128,486
24 150 39,163 5,161 229,608
25 13 10,306 171 36,874
183
TABLE D-2--(Continued)
Industry Unadjusted Fully Adjusted
Number* (77:1) ( KL) (2(4) ( 7! L)
Method VI (Continued)
26 10,080 346,897 18,023 463,856
27 895 50,453 5,202 121,642
28 2,047 88,972 5,171 141,406
29 233,986 1,937,644 261,331 2,047,737
30 20,244 362,043 27,506 422,011
31 448,629 1,919,454 479,664 1,984,735
32 28,035 72,930 29,254 74,498
33 129 46,433 804 115,764
34 2,544 135,367 4,431 178,647
35 307 124,257 4,589 480,404
36 5,632 374,814 13,980 590,510
37 18 15,110 10,385 367,962
38 0 213 3,613 145,814
39 21 10,015 6,068 169,671
40 32 28,503 2,820 268,009
41 10,922 279,353 17,854 357,164
42 906 76,003 2,645 129,876
43 4,743 247,591 14,795 437,262
44 1 2,596 1,574 115,605
45 451,661 577,402 641,308 688,027
46 514,577 752,338 654,514 848,491
47 4,639,658 534,050 4,963,004 552,346
48 19,414 129,728 50,819 209,890
49 872 93,499 1,795 134,149
50 91,756 395,806 105,300 424,012
51 1,124,368 1,759,919 1,370,948 1,943,341
52 741,081 126,381 770,789 128,889
53 4,896 98,642 11,283 149,739
54 62,385 484,113 82,549 556,881
55 20,306 311,978 27,216 361,184
56 25,092 279,351 33,913 324,762
57 6,414 91,628 14,797 139,173
58 143 14,135 2,728 61,731
59 1,413,295 3,468,735 1,532,780 3,612,390
60 9 2,333 149 9,642
Total "Welfare
Loss" 10,398,164 22,655,766 11,855,472 28,650,314
184
.._. _. __.__———-— _4_..*_
Industry Unadjusted Fully Adjusted
Number* <4“) (11.) <7(=1) (“711)
Method VII
1 56,103 959,053 79,781 1,143,665
2 13,652 120,895 8,398 94,819
3 8,969 103,723 9,430 106,359
4 9,953 171,351 12,090 188,857
5 683 26,547 1,201 35,217
6 17,790 666,883 25,524 798,804
7 350 37,714 15,113 247,800
8 1,730 204,650 19,627 689,320
9 415 30,722 7,644 131,901
10 7,764 229,000 ‘1l,832 282,706
11 196 34,677 1,546 97,408
12 652 52,814 1,347 75,928
13 80 14,242 994 50,149
14 51 16,848 914 71,324
15 1,784 107,593 4,295 166,943
16 85 32,260 21,712 515,818
17 96,593 1,323,330 112,383 1,427,403
18 0 274 1,055 60,008
19 230 20,866 1,007 43,653
20 372 42,598 1,532 86,402
21 107,651 1,205,787 120,609 1,276,291
22 199 42,973 2,191 142,726
23 62,956 909,925 85,491 1,060,344
24 657 81,948 7,224 271,666
25 17 11,537 180 37,841
26 13,473 401,045 22,391 517,011
27 1,483 64,940 6,485 135,811
28 1,403 73,651 4,473 131,522
29 194,518 1,766,685 223,177 1,892,358
30 10,393 259,413 15,672 318,543
31 349,701 1,694,659 376,088 1,757,433
32 22,738 65,679 23,799 67,195
33 88 38,265 695 107,623
34 56 20,047 531 61,836
35 122 78,219 3,740 433,719
36 1,826 213,406 7,440 430,778
37 83 32,946 8,601 334,876
38 36 14,565 4,361 160,185
39 9 6,632 5,594 162,909
40 81 45,451 3,196 285,032
41 2,094 122,331 5,928 205,807
42 1,730 105,021 3,984 159,386
43 3,283 205,968 12,519 402,233
44 18 12,360 1,242 102,709
185
TABLE D-2--(Continued)
Industry Unadjusted Fully Adjusted
Number* ()7 =1) (71 L) , (7(=1) (71 L)
45 638,701 686,627 868,312 800,590
46 373,082 640,605 493,015 736,407
47 3,467,307 461,673 3,750,185 480,137
48 20 4,125 13,890 109,730
49 150 38,748 0 1,321
50 74,890 357,582 90,876 393,903
51 2,300,762 2,517,528 2,688,736 2,721,527
52 623,092 115,884 653,778 118,704
53 6,928 117,330 13,951 166,503
54 55,024 454,653 76,827 537,233
55 19,323 304,335 32,841 396,756
56 20,587 253,034 29,112 300,898
57 5,777 86,961 15,489 142,393
58 8 3,347 2,140 54,676
59 1,337,649 3,374,627 1,499,642 3,573,127
60 512 17,865 1,030 25,334
Total ”Welfare 7
Loss” 9,915,876 21,104,421 11,512,863 27,329,557
Method VIII
1 226,086 1,925,240 242,486 1,993,847
2 6,384 82,672 7,051 86,885
3 6,803 90,333 7,204 92,959
4 17,283 225,799 19,779 241,552
5 1,669 41,507 2,319 48,925
6 29,013 851,645 38,540 981,573
7 O 642 10,577 207,305
8 4,438 327,781 27,214 811,689
9 2,970 82,218 14,693 182,866
10 4,382 172,048 7,533 225,567
11 22 11,630 900 74,315
12 3,132 115,787 4,542 139,437
13 125 17,795 1,136 53,608
14 841 68,403 2,697 122,482
15 3,107 141,994 6,253 201,441
16 5,697 264,219 45,424 746,084
17 67,009 1,102,203 80,621 1,208,979
18 573 44,230 3,144 103,587
19 53 9,983 565 32,709
20 1,860 95,202 3,958 138,877
21 72,487 989,443 82,893 1,058,084
22 1 3,542 1,158 103,775
23 46,658 783,344 69,925 958,970
24 3,141 179,125 13,267 368,153
186
TABLE D-2--(Continued)
==—_==: : 5
Industry Unadjusted Fully Adjusted
Number* <7(=1) (71 1,) (7(4) (71:)
Method VIII (Continued)
25 453 60,104 927 85,951
26 17,687 459,509 27,800 576,088
27 1,184 58,027 9,839 167,289
28 449 41,677 2,287 94,043
29 62,869 1,004,374 76,124 1,105,197
30 2,305 122,173 4,963 179,258
31 16,834 371,813 21,152 416,781
32 3,002 23,863 3,272 24,917
33 393 80,973 1,349 149,922
34 42 17,304 488 59,306
35 774 197,280 6,027 550,590
36 3,928 313,013 11,249 529,696
37 44 23,842 8,188 326,729
38 23 11,649 4,221 157,594
39 479 47,659 9,072 207,455
40 135 58,651 3,580 301,631
41 4,861 186,376 9,837 265,118
42 2,818 134,050 5,607 189,084
43 1,056 116,809 7,354 308,825
44 95 28,388 1,658 118,638
45 543,436 633,354 740,477 739,312
46 259,930 534,707 339,079 610,715
47 4,222,740 509,490 4,565,751 529,779
48 685 24,362 18,524 126,721
49 1,869 136,885 3,253 180,595
50 94,289 401,231 107,684 428,785
51 513,082 1,188,864 673,046 1,361,636
52 1,238,642 163,389 1,283,974 166,352
53 11,795 153,098 20,787 203,241
54 37,717 376,422 54,501 450,617
55 22,281 326,801 29,440 375,650
56 21,225 256,922 29,425 302,511
57 4,022 72,563 10,969 119,827
58 877 34,997 6,343 94,131
59 1,322,939 3,356,020 1,506,480 3,581,264
60 738 21,454 1,332 28,821
Total "Welfare
Loss" 8,919,432 19,174,880 10,299,487 25,297,199
*See Table 7 for an outline of these different profit methods as
well as Appendix A's section "Profit Rate and'Welfare Loss'Adjustments"
for the distinction between Methods I-VIII and 1-8; also see Table 9
for Industry Number Code.
SOURCE: See Table B-1.
APPENDIX E
TWO-DIGIT VALUE-ADDED AND EMPLOYMENT CONCENTRATION RATIOS
BASED UPON PERCENTAGES ACCOUNTED FOR BY 4, 8, AND 50
LARGEST FIRMS IN AMERICAN MANUFACTURING, 1958
187
188
APPENDIX E
TWO-DIGIT VALUE-ADDED AND EMPLOYMENT CONCENTRATION RATIOS
BASED UPON PERCENTAGES ACCOUNTED FOR BY 4, 8, AND 50
LARGEST FIRMS IN AMERICAN MANUFACTURING, 1958
The concentration ratios for the various industries utilized
above and shown below are based upon concentration ratios calculated
for finer industrial subdivisions. In moving from four-digit manu-
facturing (443 industries for E, 446 for VA) to two-digit (20 indus-
tries), there are a number of problems in combining the industries.
The interested reader should consult the brief but enlightening
Appendix C in George J. Stigler's Capital and Rates of Return in Manu-
facturing:lndustries (National Bureau of Economic Research, 1963), pp.
206-215 for a description of some of the problems. To obtain our
estimates we merely averaged the concentration ratios of the shipments
(or employment) at the four-digit level with value-added (or employ-
ment) as weights. This will cause an overestimation of the correct
concentration whenever the four-digit industries are highly competi-
tive with one another. To illustrate our estimation procedure, let
us take the following hypothetical and unrealistic example of a two—
digit industry composed of two four-digit industries:
1958
Product Value of Shipments Concentration Ratio
9950 $ 80,000 40
995 20,000 90
99 100,000 ?
189
The weighted average is .80 x 40 + .20 x 90 = 32 + 18 = 50. If the
reader is interested in some kind of crude notion of what might consti-
tute an appropriate figure for concentrated vs. unconcentrated indus-
tries, Stigler's benchmarks for three-digit industries are worth
remembering. His criteria were: concentrated, if four largest firms
ship over 60% of the product in the national market; unconcentrated,
if over 50% in the national market, or under 20% in a regional market;
ambiguous (not labeled) if outside both categories. His characteriza-
tion of industries as national, regional, or local was based, with but
slight modifications, on the National Resources Committee report on
The Structure of the American Economy, Part 1, Appendix 8. The excel-
lent source for our estimates is contained in the note to the following
table containing our estimated two-digit ratios.
190
Om mm mm mm Hm 6w N4 mm ucmanfisoo :ofiumuuoamamua 4m
ow 46 mm N6 em 64 06 m6 448445068 460444064m om
N4 an 66 mm 64 mo w6 on 4460444
10640 unmuxmv humcwnumz mm
66 46 44 44 66 46 mm 44 64666644 46466 6646644664 64
mm N4 on M6 mm 64 Ho 46 64666644 46408 4468444 mm
N4 46 66 mm N4 N6 m6 mm 64066044
mmmaw van 44640 4acoum mm
66 46 on 44 46 66 mm 64 64666644
4054664 cam nonumma 4m
m4 44 mm 66 6m 44 no mm 64036044 nonnsm on
oa m4 on mm 6m «w on 6m 64666044 4606 6:6 836404464 mm
66 64 66 46 66 64 46 66 64666644
664446 can 646648650 mm
66 an «N m4 0m 4m mm ma wcwnm44434 can wc4ucwum 44
66 mm mm 64 64 46 mm 64 64666644 664446 666 46464 64
46 mm N4 m4 Hm 4m om m4 66434x4m can 643444424 mm
46 44 64 6 66 64 44 44 4646446464 4466666
muosvoua can Hanson 6N
46 44 44 44 46 mm 64 64 64666644 6444664 66466
ufiunmm uozuo vcm Hmummm< mm
N4 6m on 6N 64 mm 46 mm 64066044 4448 6444x69 mm
mm mm am an mm 4m m4 64 604640646465 ouuwnoa 4N
6m 66 66 64 46 66 66 mm 64666644 6646644 666 6664 64
664464800 «04:64800 wcwmsouu 6600 40444
umowqu ummmuwn umowumq umowumq umowumq ummwqu ummwqu uwuwumq >4umavcH :04mwmmmao
on 04 m 6 cm on w 6 HwHHumsvcH
vumvcmum
1143 you name you vmuasooo<
kucsooo< ucmamoHQBm «o 4:60 yum mucmaawnm mo 6:46> no 4460 464
Ill l.i lllyl... I. l l 1, ll 11 ,I 11.! 1.!1. l." lul'll‘ lit"
z HHUHouoBH
Hum mAm00 ".0 .4 4404w444663v 44 046 H 444mm 464646m 66464m 46444:
4446404040 644 40 664448500 644 40 4404040: 046 464444444 40 66444580044m 644 404 646460 644 40 46644m
644 44 06464644 440464 .wmm4 4446:446Iwm44440644462 44 404464 4044644460400 8044 06444800 4004500
466.46 6 64 446.66 6 6>
464.66 a 44 466.66 6 64
446.66 6 66 446.46 6 44 x
Nwm.o6 u m> N6N.o6 u 4> "646 .HWW n > 440446446> mo 644640444600 64H
mm N6 mN 0N 4m M6 4n NN 66444
10644468 640646446064z mm
46 64 66 66 66 64 66 66 64666644
0646464 046 64468444644 mm
464464600 664464800 w4444040 6400 40446
466w464 4460464 4660464 4660464 466w464 466w464 446w464 466w464 44464444 1044466640
on ON 0 6 on ON 0 6 4644464444
0460464m
uumn 40w 1144 now 06444000<
06444000< 4468404430 40 4460 46m 64468444m 40 6446> 40 4460 46m
46666446604--4-4 44664
APPENDIX F
RANKING OF INDUSTRIES BY LERNER"S INDEX OF
MONOPOLY POWER, Zm = (P-MC)/P
192
193
TABLE F-l
RANKING 0F INDUSTRIES BY LERNER'S INDEX OF
MONOPOLY POWER, Zm = (P-MC)/P*
Using Using
Profit Method I, Fully Profit Method IV, Fully
Adjusted,with 72: 1 Adjusted, with 77: 1
, Industry Industr
RankIng** Number*** Number*i*
l 45 59
2 59 50
3 50 48
4 54 54
5 56 56
6 55 55
7 60 60
8 9 58
9 41 9
10 16 16
11 58 41
12 7 6
13 25 7
14 6 25
15 39 39
16 43 43
17 38 38
18 46 27
19 27 36
20 49 37
21 37 8
22 36 15
23 26 26
24 15 24
25 8 44
26 24 49
27 44 40
28 13 18
29 35 13
3O 4O 35
31 18 33
32 11 11
33 33 34
34 20 23
35 5 20
36 23 ' 19
37 . 34 42
38 42 14
194
TABLE F-l--(Continued)
Using Using
Profit Method I, Fully Profit Method IV, Fully
Adjusted, with 71= 1 Adjusted, with 71: 1
Industry Industry
Ranking** Number*“ Numberm
39 19 2
4O 22 22
41 28 57
42 14 21
43 12 12
44 10 10
45 21 30
46 57 53
47 48 5
48 3O 17
49 17 1
50 2 29
51 53 3
52 l 31
53 29 4
54 4 2
55 2 32
56 31 51
57 32 46
58 51 45
59 47 47
6O 52 52
*v
*Computed from profit Methods I and IV outlined in Table 7
and discussed in Appendix A, "Profit Rate and'Welfare Los§.Adjustments."
**Rankings are from highest monopoly power to lowest;
***For Industry Number coding see Table 6.
SOURQE: See Table B-1.