On Merit Goods

Author(s): John G. Head

Source: FinanzArchiv / Public Finance Analysis , 1966, New Series, Bd. 25, H. 1 (1966),
pp. 1-29
Published by: Mohr Siebeck GmbH & Co. KG

Stable URL: https://www.jstor.org/stable/40909986

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Mohr Siebeck GmbH & Co. KG is collaborating with JSTOR to digitize, preserve and extend
access to FinanzArchiv / Public Finance Analysis

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods*
by

John G. Head

1. Introduction

Professor Musgrave' s great treatise "The Theory of Public Finance"1 is

based on the welfare foundations provided by his "multiple theory of the
public household." 2 Apart from providing much the most elegant version of a
post-Keynesian welfare framework familiar from the public finance literature
of the post-war period,3 the central contribution of Musgrave' s multiple
theory is to be found in the twin concepts of "social wants" and "merit
wants" which dominate his discussion of the Allocation Branch.4 Of these,
most attention has so far been paid to the social wants concept which, on
examination, proves to be identical with the concept of a public good which
has figured so prominently in Continental discussions. Following the pioneer-
ing work of Baumöl, it has been shown that the public good concept, although
of little significance in itself even in relation to allocation problems, can read-
ily be generalised into Sidgwick's concept of non-appropriability to become

* The research for this paper was carried out while the author was a Visiting
Fulbright Scholar at Columbia and Princeton Universities in the United States
during the academic year 1964-65. He is particularly indebted to Richard Musgrave
and Carl Shoup for helpful discussions.
1 R. A. Musgrave: The Theory of Public Finance, New York 1959. (Musgrave^
book is now available in a German translation by Lore Kullmer.)
* Musgrave, ibid., un. i. - ¿leaders ot this journal will be laminar with tne ear-
lier version of this theory contained in his article "A Multiple Theory of Budget
Determination", Finanzarchiv, N. F. Bd. 17, 1956-57. See also his paper "Principles
of Budget Determination", in: Joint Economic Committee, Federal Expenditure
Policy for Economic Growth and Stability, Washington 1957.
3 For a general discussion of this post-Keynesian framework and its relationship
to modern welfare economics, see J.O.Head: The Welfare Foundations of Public
Finance Theory, Rivista di diritto finanziario e scienza delle finanze, 1965, especially
Section 4.
4 Musgrave, op. cit., (Jh. 1, bection B. It is interesting to notice that the merit
wants concept appears to have been developed rather late in Musgrave's thinking on
the "multiple theory". No such concept is mentioned in his Finanzarchiv article,
and only one rather cryptic paragraph is devoted to it in his Joint Economic Com-
mittee paper (p. 111).

1 Finanzarchiv N. F. 25 Heft 1

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 2 John O. Head

an important contributing f
of resources, economic instab
The considerable interest which has been shown in the clarification and
further development of the social wants concept stands, however, in sharp
contrast to the comparative neglect of merit wants. It might perhaps be
thought that this neglect simply reflects the relative simplicity or unimpor-
tance of the merit wants concept, or the completeness of Musgrave's own
discussion. On close examination, however, the apparent simplicity of the
concept soon disappears, and its importance becomes increasingly obvious.
Similarly, as in the case of social wants, Musgrave's own discussion, though
extremely interesting and stimulating, is somewhat unclear, and he makes
no attempt to relate the concept to more familiar theories of public policy.
It is therefore the aim of the present article to examine the meaning of the
merit wants concept as it appears in Musgrave's theory, and to relate it to
more familiar Pigovian, Keynesian and post-Keynesian theories of public
policy. In this way it is hoped to contribute to the clarification and further
development of this important but rather enigmatic public finance concept.
We shall begin (Section 2) with a brief outline of Musgrave's merit want
theory. Then (Section 3) we shall set out and examine in detail the main cha-
racteristics of merit goods. In Section 4 we shall attempt to relate a generalised
concept of merit goods to other important concepts in modern theories of
market failure. In Section 5 we shall compare the resulting merit goods
theory of market failure with the generalised public goods theory which has
emerged from recent discussions. The political problem of improving on mar-
ket performance in the case of merit goods will be discussed in Section 6.
Section 7 contains a brief summary of the main conclusions.

2. Musgrave's Theory of Merit Wants

In the case of social wants or public goods, the need for public policy
arises from the failure of the market to allocate resources in accordance with
given individual preferences. 2 In contrast, in the case of merit wants, the
preference maps of individuals are no longer taken as given, but are themsel-
ves subject to critical scrutiny. Public policy in this area is justified by the
need to correct individual preferences. As Musgrave expresses it, "...the
satisfaction of social wants falls within the realm of consumer sovereignty,
as does the satisfaction of private wants. The satisfaction of merit wants, by
its very nature, involves interference with consumer preferences."3
1 For a detailed discussion and further references, see J.O.Head: Public Goods
and Public Policy, in: Public Finance/Finances Publiques, 1962.
2 It is easy to show that pure public goods, defined as those which must be
consumed in equal amounts by all, exhibit to an extreme degree special forms of two
conceptually quite distinct and independent characteristics, namely joint supply
and external economies. As a result, it is impossible for decentralised market insti-
tutions to achieve the Parefo-optimum based on the existing pattern of factor
ownership, technology and individual preferences. See Heady ibid.
8 Musgrave, op. cit., p. 16.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 3

Since merit goods are those sat

defined as those of which, due to
consume too little. In such cases
consumption. As possible examp
wants, Musgrave mentions pub
low-cost housing and free educ
be defined as those of which, due
to consume too much. Here the
consumption. Musgrave sugge
such intervention.2
To justify interference with consumer sovereignty for the purpose of
satisfying merit (or demerit) wants, and thus to justify the place of the merit
wants concept in his welfare framework, Musgrave offers three arguments.3
According to the first, "wants that appear to be merit wants may involve
substantial elements of social wants". Educational services and health meas-
ures are mentioned as examples. Secondly, under certain conditions, an "in-
formed group" may be justified in "imposing its decision upon others". Ex-
amples suggested include education, health services, protection of the inter-
ests of minors, avoidance of undesirable discrimination, and control of the
sale of drugs. Finally, due to persuasive advertising, "the ideal of consumer
sovereignty and the reality of consumer choice in high-pressure markets may
be quite different things".

3. The Properties of Merit Goods

Turning now to more detailed consideration of the merit goods concept,

it appears that merit goods may raise as many as three distinct problems,
namely preference-distortion problems, distributional problems, and public
goods problems.
3.1 Preference-distortion Problems

It is evident from Musgrave's discussion that distorted preferences con-

stitute the essence of the merit goods problem ; individuals find the benefits from
merit goods particularly difficult to evaluate. By contrast, the most striking
common characteristic of modern welfare theories is the uncritical accep-
tance of given individual preferences as the foundation of welfare economics ;
and the rather vague notion of consumer sovereignty has been given precise
expression in the crucial concept of a PareJo-optimum. 4 The critical attitude
to individual preferences which is the central feature of the merit wants
approach thus represents a striking departure from the modern welfare tra-
dition.

1 Idem.
2 Idem.
3 Ibid., pp. 13-14.
4 In its most general form, the "Social Welfare Function" version of modern
welfare theory allows, at least conceptually, for the possibility of rejecting or modi-

l*

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 4 John 0. Head

3.11 Preference-distortion A
critical attitude to individua
literature. An interesting c
between productive and unp
clude Marshall1 s well-known classification of wants into the natural and the
artificial, and the more systematic criticisms of the given wants approach
to be found in the works of J. M. Clark and F. H. Knight, * In the more recent
literature the critical attitude to the given wants approach has been very
forcefully represented by such writers as Galbraith and Scitovsky. 2
In traditional public finance discussions, preference-distortion arguments
for government intervention are commonplace. Typical examples concern
the treatment of minors and lunatics, lower class education, drug addiction,
and the choice between luxuries and necessities. 3 Indeed it is hardly too much
to say that arguments regarding erroneous individual preferences and the
superior knowledge of some elite group or government dominate public
policy discussions in most of the older literature. Perhaps the most celebrated
example of a merit goods discussion to be found in the public finance liter-
ature is, however, Pigou's argument regarding the irrational discount which
individuals place on future as compared with present consumption and the
consequent bias against capital accumulation and growth.4
3.12 The Causes of Distorted Preferences. - What are the main causes of dis-
torted preferences ? From the diversity of Musgrave's own examples and
others to be found in the literature, it appears that there are important dis-
tinctions to be drawn between the different types of imperfect knowledge
which may be involved. In particular, it seems desirable to distinguish be-
tween preference distortions due to uncertainty (i. e. incomplete or inaccurate
information) and those due to irrationality.
3.121 Uncertainty. - The overwhelming importance of uncertainty in individ-
ual behaviour has often been noted. Individual preferences for most goods
are inevitably based on incomplete information, the degree of incompleteness
varying with such factors as the complexity of the item, the indirectness or
remoteness of the benefits, and the frequency of purchases. Additional in-
formation is costly to acquire, and a substantial degree of ignorance will
therefore usually be completely rational. The difficulties facing the individual

fying individual preferences. Nevertheless, practitioners of this approach have

usually accepted individual preferences without serious question. For a very sti-
mulating discussion of the problem by a leading welfare theorist, see, however,
T. Scitovsky: Papers on Welfare and Growth, London 1964, Part C. Also J.Rothen-
berg: Consumers' Sovereignty Revisited and the Hospitability of Freedom of Choice,
in: American Economic Review, Papers and Proceedings, 1962.
1 For an excelent survey and analysis of the earlier literature on this subject,
see H.Myint: Theories of Welfare Economics, London 1948, Ch. XL
2 See, for example, J.K.Galbraith: The Affluent Society, .Boston 1958, and
T. Scitovsky, op. cit., Part C.
3 Such examples figure prominently even in Mills classic liberal analysis of
the economic functions of government. See J.S.Mill: Principles of Political Eco-
nomy, Book V.
4 A.C.Pigou: A Study in Public Finance, 1st edition, London 1928, fart 11,
Ch. VIII, Sections

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 5

in the purchase of the modern a

tially dangerous consumer durab
problem. A more traditional exam
informed to the benefits of edu
plete information is also a crucia
Moreover, much information
persuasive and misleading rather
most types of advertising. Galbra
public sector products due to hi
merely provides the most graphi
al argument.
In all these cases it may be pos
advice of technical experts or ot
the most important uncertainti
mation, or otherwise to correct th
sible examples of such governmen
apprenticeship and registration
trades and professions, various ty
ling the sale of dangerous drugs, s
programs, sumptuary excises, an
Experience of increased or red
particular products as a result o
here tend to result in the eventu
tures. The individual should himse
of his pre-intervention choices. T
consumption of a particular goo
a pure comparative statics fram
goods policies of this sort should
In a dynamic world of new co
goods problems must, however
therefore have to be retained and new measures introduced.
The case for regarding the correction of preference distortions due to
uncertainty as a legitimate economic function of government is obviously a
strong one. Indeed such functions can scarcely be regarded as lying outside
the consumer sovereignty tradition. The acceptance of actual choices or
"revealed preferences", however remote from true preferences, clearly re-
quires a value judgment at least as controversial as the merit goods approach.

3.122 Irrationality. - Even where information is complete and accurate,

however, a concept of distorted preferences is still possible.
Thus it is conceivable that an individual may make an erroneous choice
even when he is in possession of all the relevant information. Such a choice we
may define as irrational. Obvious examples are provided by the preferences of
the psychologically disturbed and the immature. In these cases, preferences for
a wide range of goods may be involved, and the irrationality arises less from
the inherent nature of the good than from the particular class of consumer.
The habit-forming drugs, however, provide an example where the irrational-

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
6 John G. Head

ity tends to be associated mo

In any actual case, including
ely difficult to disentangle
correct information as a sour
Apart from the somewhat
pure irrationality poses obv
sidering Pigou's argument re
consumption, which is much
rationality argument which
adults in the public finance
everybody prefers present p
future pleasures or satisfact
perfectly certain to occur. Bu
the idea is self-contradictory
tude is any greater than a fu
only that our telescopic facul
pleasures, as it were, on a d
mation and economic growt
nomic function of governme
In its pure form as stated b
quite appealing, has proved
criticized as authoritarian an
really prefers present to fut
This is an inevitable welfare
pealing for its moral author
on closer examination, much
turns out to derive from the
temporal choices.
The concept of erroneous p
information is therefore fr
cannot be avoided by appeal
group or ethical observer as
quite rightly) what practitio
cize.3 In spite of these diffi
however, it could still quite
of the "true preferences" of
comprehensive welfare fram
1 Ibid., p. 117. (First set of ita
in The Economics of Welfare,
that individuals prefer prese
earlier by such writers as Men
ticular importance for its emp
2 See, for example, Myint, op
3 See, for example, Scitovsky,
a majority decision in this area
argument, see Erik Lindahl: T
mic Papers, No. 10, London an
4 Welfare theories, by their
proved" in any usual scientific

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 7

3.2 Distributional Problems

Although distorted preferences are clearly the essential characteristic of

Musgrave's merit goods concept, his discussion and choice of examples sug-
gest the existence of further characteristics. Thus the examples of low-cost
housing, education, school luncheons, and health services all suggest that
important distributional problems may be involved.
This impression is confirmed in the course of his discussion of distribution,
where he observes that" . . . the satisfaction of merit wants is associated frequently
with distributional considerations. Subsidies in kind, for instance, may be given
for the satisfaction of certain private wants. Free charity clinics may be furnished,
or subsidies may be given to low-cost housing, the benefits of which will accrue to
low-income families. Such programs have a dual nature. They may be looked upon
as operations of the Allocation Branch in the satisfaction of merit wants and as
operations of the Distribution Branch in the redistribution of income." 1 In his later
discussion of public sales, he analyses the case of low-cost housing in rather more
detail as follows: " . . . the government may wish to sell at a loss ... to subsidize the
buyer of particular products. Such is the case with public operations in low- cost
housing. ...In the efficient system, a loss operation in low-cost housing involves
(1) a policy of distributional adjustment in favor of low incomes, since only low-cost
housing is involved; (2) recognition of housing as a merit want, since the distribution-
al adjustment is not made in cash. . ." 2
If, as suggested in the above argument, housing in general is a merit
good, one might wonder why only low-cost housing is mentioned in Musgrave's
merit want discussion. Again, a loss operation in low-cost housing may be
allocatively efficient, but in terms of Musgrave's own familiar conception of the
proper state of distribution in terms of horizontal and vertical equity it
appears that the operation must be distributionally inefficient. In particular,
the programme discriminates between different low-income individuals ac-
cording to their expenditures on housing, and significant horizontal inequities
are therefore practically unavoidable. For horizontal equity in the conventio-
nal sense, the distributional adjustment would have to be made in cash, even
where distorted preferences are involved. The cost of the Allocation Branch
operation in low-cost housing should be charged to the low-income individ-
uals concerned, and the purchase made compulsory. To avoid the apparent
inconsistency in Musgrave's discussion, the obvious conclusion would seem
to be that in a number of significant cases merit goods pose distributional
problems of a rather special sort; in these cases the distributional objective is
defined partly in terms of the distribution of these particular goods, and not
merely in terms of "income".
Although this may not have been Musgrave's explicit intention, such a
conception has considerable merits. Conceptions of equity or the proper state
of distribution seldom run solely in terms of Lorenz-cuives of income. Indeed,
in relatively conservative democracies such as the United States, distribution-
al objectives frequently seem to be formulated in terms of particular goods

and their „validity" must therefore depend upon the persuasiveness of the argu-
ments in terms of general acceptability, community advantage, social justice, objec-
tive rationality, etc. which can be mustered in support.
1 Musgrave, op. cit., p. 21.
2 Ibid., p. 49.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 8 John G. Head

rather than income.1 Thus t

of equality (or at least high
health services, even where
finance concepts of horizo
Lorenz-cuTves of income. Th
strong arguments regarding
tion problems and consequen
amongst the lower income g
grave' s merit goods concept
Is there, however, any nece
ences and distributional obje
it seems clear that the two c
need be in no way related. D
which pose no particular pr
housing project would appea
so-called sumptuary excise
provide examples where no
consideration here (or indee
therefore of the greatest im
these characteristics.

3.3 Public Goods Problems

The third important characteristic of the merit goods concept is that

public goods (or social wants) problems appear to be involved. This is evident
from the following passage containing the first of the three arguments which
Musgrave puts forward to justify the place of the merit goods concept in his
normative framework: "To begin with, situations arise that seem to involve
merit wants but on closer inspection involve social wants. Certain public
wants may fall on the border line between private and social wants, where the
exclusion principle can be applied to part of the benefits gained but not to all.
Budgetary provision for free educational services or for free health measures
are cases in point. Such measures are of immediate benefit to the particular
pupil or patient, but apart from this, everyone stands to gain from living in
a more educated or healthier community. Wants that appear to be merit
wants may involve substantial elements of social wants."2 Since this argu-
ment is presumably intended to justify the merit wants rather than the social
wants concept, it seems reasonable to conclude that merit goods typically
exhibit public goods characteristics.
Although once again Musgrave's intentions are not entirely clear from his
discussion, the idea that goods with merit aspects also typically, or at least
frequently, exhibit public aspects has much to commend it, as his own merit
goods examples of health and educational services clearly show. The existence
1 Compare Sdtovsky^ recent suggestion (op. cit., Ch. 16) that equity is mainly
a question of the distribution of "necessities". More generally, it appears that, in
conceptions of equity, different weights are assigned to the distribution of different
classes of commodities.
£ Musgrave, op. cit., p. lá.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 9

of such a relationship has in fac

goods literature. Indeed, the ve
usually been held to be not the
teristics emphasized in recent A
sibility of individual evaluation o
goods have been said to be too i
susceptible of accurate appraisa
duals. In these matters individu
the bureaucracy or government
democratic versions of the sam
consideration of such difficult m
many goods pose both preferenc
a concept of merit goods emb
attention on a most important
Once again, however, it is extr
is any necessary relationship be
problems. It has recently been s
which must be consumed in equ
degree special forms of two dist
joint supply and external econo
ever, is either implied by, or im
luation problems. Thus classic e
formance demonstrate clearly
indivisibility of product in the
any obvious problems of benefi
fireworks display, most individ
benefits involved, in spite of th
price-exclusion problems. In the
formation, misleading advertis
severely distorted preferences
goods. Thus although recent An
as that of Samuelson, probably
whelming problems facing the
he is likely to derive from some
ditional identification of public
problems is clearly incorrect an
lems are conceptually quite dist
From the above analysis it
three conceptually quite distinc
dual evaluation problems, distri
1 For a detailed discussion, see
(loc. cit.).
¿ Jbor a celebrated statement ot this argument, see Antonio de Viti de Marco:
First Principles of Public Finance, London 1936, p. 118, footnote 1.
3 These difficulties have not, however, been ignored by all writers in the in-
dividualistic tradition. Such great pioneers of this approach as Sax, De Viti, Wicksell
and Lindahl all concede that public goods frequently pose extremely difficult indi-
vidual benefit-evaluation problems.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 10 John G. Head

Although there is no necessary

goods concept is extremely us
very significant coincidence th
formulation problems frequen
teristics and at the same tim
problems of distribution.

4. The Relationship between

Theory of Public Po

Let us now consider more ge

racteristics of the merit goods
dern theories of market failu
it appears that Musgrave's me
able and very fruitful general

4.1 Distribution

We have seen that, although merit goods pose special redistributive prob-
lems, there is no necessary entailment relationship between the special prob-
lem of distributional objectives applied to particular goods and the individ-
ual benefit-evaluation problems which are the essence of the merit goods
concept. It appears, however, that there are extremely important relation-
ships between distributional objectives and the generalised problem of im-
perfect knowledge.
4.11 Uncertainty and Income-Insurance Measures. - Thus individual uncer-
tainty concerning the future may play an important role in the formulation
of conceptions of the proper state of distribution and the need for redistri-
butive measures. The individual's own future income, not to mention that of
his family, is usually subject to a considerable degree of uncertainty. Even in
relatively stable societies, incomes of individuals of the same age, and even
of similar educational background and other socio-economic characteristics,
show considerable dispersion. 1 In societies which have been ravaged by great
wars, revolutions, depressions, hyper-inflations, and other political or social
upheavals, such uncertainty may clearly be intense. At least to a substantial
degree, such income-uncertainty is likely to be privately unavoidable and un-
insurable; and in a society in which most individuals are risk-averters, this
uncertainty may be reflected in redistributive government policies and mildly
egalitarian conceptions of the proper state of distribution, either of income in
general or of particularly important commodities (as in the case of merit
goods). Thus an economic policy objective of a proper state of distribution
may to some extent reflect individual attitudes towards privately unavoid-
able and uninsurable risk.

1 For the United States, see, for example, J.Morgan: The Anatomy of Income
Distribution, in: Review of Economics and Statistics, 1962.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 11

The government can perform its

two main ways, either by acting 'p
paratively by providing some for
appropriability problems involvin
the individual himself is unable to
where voluntary cooperative actio
unstable, preventive governme
Economic stabilisation policies of a
tive activities affecting income
appear to be a good example in t
education is an example with impor
of income and for the distribution
that such activities may be desira
ments exist, since the latter may
Hence the case for such activities a
The second obvious method is the
At least beyond a point, preventiv
risk may be impossible or relative
adequate, the promotion of more
attempted. By providing appropri
regulating private insurance arran
increase the risk-pooling facilities
time, perhaps by making member
nomies of scale not normally poss
cial security schemes, such as old
ment "insurance", provide obvio
governmental income-insurance. M
commodities also possess significan
sive income taxation may be justif
Medical insurance schemes, includ
or less generous deductions for m
excellent examples for a particu
quoted it is clear that these public
resemblance to private insurance.2
the public arrangements. On the c
zed departures of "social insuranc
insurance principles can be regard
range of risks which can be pooled
It is obvious that action to reduc
action to promote more adequate

1 For an excellent recent discussion

medical care, see K.J. Arrow: Uncert
Care, in: American Economic Review, 1963.
2 On the departures from private insurance principles in the field of social se-
curity, and the largely public relations function of the term "insurance" in this con-
text, see the classic study by A. T. Peacock: The Economics of National Insurance,
London 1952, especially Ch. III.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 12 John 0. Head

able extent substitutes. Thus,

substitute for unemployment
health measures may substitu
the two approaches should, of
As in the corresponding but
individual benefit-evaluation d
function of government is cle
consumer sovereignty traditi
ducing the dispersion of indiv
particular distributional signi
essentially one of reducing the
categories of purchases.
The insurance function of go
in the public finance literatur
theorists.1 The relation to the
until recently been largely ov

4.12 Ethical Authoritarianism

Many of the best-known stat

economic function of governm
attitudes to risk. In the Anglo
viously true of the powerful
theories of equitable taxatio
equally true of Henry Simons1
case for progressive taxation,
gression in taxation must be
ethical or aesthetic judgment
income reveals a degree (and/o
unlovely."5 In these and ma
strong element of ethical auth
ing it, the distributional valu
vious, requiring only to be cl
or irrational.
There is clearly a close relationship between this tradition of ethical
authoritarianism in discussions of distribution, and the irrationality concept
in the merit goods discussion. In both cases there is the same appeal to the

1 From a vast literature, see, for example, A.Thiers: De la Propriété, 1848.

2 A remarkable exception is to be iound m Edgeworth & celebrated argument tor
utilitarian welfare objectives in terms of expected utility maximisation. See F. Y.
Edgeworth: The Pure Theory of Taxation, Economic Journal, 1897, reprinted in:
R.A.Musgrave and A. T. Peacock (eds.): Classics in the Theory of Public Finance,
London and New York 1958, pp. 120-121.
3 In the more recent literature, the relationship between uncertainty and re-
distributive measures has been particularly emphasized in J.M.Buchanan and
O.Tullock: The Calculus of Consent, Ann Arbor 1962, Ch. 13.
4 With the notable exception of Edgeworth. See footnote 2 above.
6 H.C.Simons: Personal Income Taxation, Chicago 1938, pp. 18-19.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 13

authority of experts, elites, and in

preferences or even majority prefer
sense, equity may be said to provide
authoritarian type. In this connecti
Pigou's famous argument regarding
extremely important (intergenera
distribution is clearly not a matter
by individual preferences.
The welfare weakness of the auth
jective of economic policy is the sam
for merit goods. Any value judgment
authority of some enlightened elite
versial, and the resulting theory of p
influence action and opinion. This i
general abandonment of utilitarian
of modern welfare theories. As in
however, it is still possible to argu
preferences, including attitudes to
adequate welfare framework for pub
of equity. 1
4.2 Allocation

We have seen that, although some of the best examples of public goods
pose extremely difficult preference-formulation problems, there is no neces-
sary relationship between preference-formulation problems and the special
forms of joint supply and external economies which characterise pure public
goods.
It is easy to show that Samuelsonian joint supply (or indivisibility of
product) accounts for a particular category of decreasing cost problems,2
which in turn represent a special but extremely important case within the
general category of problems of imperfect competition. Similarly, the exter-
nal economies or extreme exclusion problems characteristic of a pure public
good are simply a special case of the general problem of external economies
and diseconomies. External economies and diseconomies, and imperfections
of competition, including those due to decreasing costs, together constitute
the major concepts in the traditional Pigovian theory of market failure in
the field of resource allocation. The Pigovian theory can therefore be regarded
as a generalised public goods theory.
Are there not, however, extremely important relationships of a more
general sort between these crucial Pigovian inefficiency concepts and the
problems of imperfect knowledge ?
4.21 Uncertainty and Imperfect Competition. - The first of these relationships
concerns the role of uncertainty in promoting and preserving imperfections
1 For a critique of modern welfare theory along these lines, see J.G.Head: The
Welfare Foundations of Public Finance Theory (loc. cit.), Section 2.4. See also foot-
note 4, p. 6 above.
2 See J.G.Head: Public Goods and Public Policy (loc. cit.), Section 4.21.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 14 John G. Head

of competition. x Thus where, fo

good makes it difficult for the
ences, the incentive to effectiv
ficantly reduced. Moreover, un
to regard price, along with the
portant index of quality. 2 Sin
indication of inferior quality
incentive to price competition
At the same time, uncertaint
tive advertising and superfi
models" as weapons of competi
sources and misleading consum
returns to scale, thus operatin
firms. Similarly, due to decrea
vice arrangements as a form o
nically complex consumer dura
trants. The most important un
is the ignorant consumer's habi
by the size, age and reputation
purchase existing goodwill, or c
tising campaign. This same hab
market imperfections which pa
In short, by restricting quali
firms and by obstructing the e
prime cause of imperfect comp
4.22 Uncertainty, External Econ
but also extremely important r
vian inefficiency concepts can
above where it appears that ma
be traced to the existence of
particular, due to generalised p
may find themselves unable, ei
ation, to reduce a certain type
ations may indicate that the
partially non-excludable form
This would certainly seem to b
vices as defence, public health
where the risk is not narrowly

1 The following analysis is base

Ch. 12).
* Jror a penetrating analysis ot the nabit oí judging quality Dy price, see
Scitovslcy, op. cit., Ch. 11.
3 tScitovsky (op. cit., p. 207) has pointedly observed that established hrms
"...often pursue an advertising policy that seems deliberately aimed at impressing
the consumer with his own ignorance. . . . All such advertising carries the suggestion
that the consumer, a mere layman, would be unwise to judge quality unaided, by
mere inspection, and should rely instead on the guarantees offered by the reputation
of established manufacturers." See also ibid., p. 194.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 15

commodities. Similarly, where prev

the risk is too costly and private in
government may be able to provi
facilities whilst at the same time ac
technically possible but often not p
government intervention. Health in
In our discussion of merit goods in
also considered the general category
chase of particular commodities, w
for such measures as standards and
course, many forms of risk-reduci
commodities can be provided quite
only in so far as the risk-reducing o
partially non-excludable form and/
that a prima facie case for governm
the case of the general risk-reducin
can be little doubt that these Pigovia
important in the field of particular
cooperative action by private econom
4.23 Irrationality and External Ec
irrationality concept there is an im
concept of external economies. Thu
by "informed groups" to the true p
simply a special form of the gener
phenomenon. In many of the exampl
of less educated groups, it is easy to
by a sophisticated elite regularly ex
congenial popular culture. Even wher
forward, as in the case of scholarly
tions of the proper state of distrib
elite such as a priestly caste substan
munity, it is nevertheless quite clea
functions of other individuals enter

1 See Arrow, op. cit., p. 963.

¿ Tims, tor example, guarantees (ot varying duration) and service contracts are
commonplace in the field of the more expensive and technically complex consumer
durables. Likewise it is obvious that many safety features and quality controls re-
quire no legislation, but are adopted in response to market pressures. Similarly,
voluntary cooperation in such ventures as consumer associations can also do much
to reduce uncertainty in an otherwise inexpert market.
3 It should be noted that much advertising and public relations activity re-
duces subjective uncertainty without in any way affecting objective uncertainty. The
"special personal relationship" established with customers (or patients in the case
of the medical profession), now largely replaced by advertising and "brand images"
in most fields, falls in this category.
* Une very spe.cial example particularly emphasized by ¡Scitovslcy (op. cit.,
Chs. 13 and 15) is the threat to the traditional educational function of elite tastes
represented inter alia by the overwhelming importance of economies of scale under
modern conditions.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 16 John O. Head

tions of the informed group. T

treated as a special case of ex
appears that at least some part
can in fact be reconciled with
tradition.
To suggest that the concept of irrational preferences can be completely
explained in terms of external economies and diseconomies would, however,
be extremely misleading. In particular, there is always a strong implication
in irrationality arguments that the judgments concerned are in some sense
completely objective and disinterested. In such cases the external economy
or diseconomy element is by definition very weak or non-existent ; or the same
point might alternatively be expressed by saying that externalities of this
particular sort are defined to be Parefo-irrelevant. Strictly defined in this way,
the irrationality concept remains completely authoritarian and is irreconcil-
able with the individualistic tradition.

4.3 Economic Stability

There are also obvious and important relationships between uncertainty

and the Keynesian and post-Keynesian theories of unemployment and in-
flation. According to orthodox Keynesian theory, crucial factors contributing
to the instability of the economic system include interest-elasticity of the
liquidity-preference schedule, interest-inelasticity of the investment-demand
schedule, and elastic price and interest expectations. * As emphasized by Key-
nes himself, these factors are essentially phenomena of a highly risky and un-
certain world.

Thus the role of uncertainty in the theory of liquidity preference is immedia-

tely evident from Keynes* discussions of the precautionary and speculative demands
(and of course from the terms themselves).2 The precautionary demand is analysed
in terms of risk-aversion, and the speculative demand in terms of differing opinions
held with subjective certainty. A major feature of more recent discussions has been
the increasingly rigorous analysis of liquidity-preference as behaviour towards
risk. 8
Similarly, the interest-inelasticity of the investment -demand schedule has
been explained largely in terms of the overwhelming importance of risk in investment
decisions, and the consequent relative unimportance of feasible changes in rates of
interest compared with the large risk premium demanded by investors. The pheno-
mena of elastic price and interest expectations along with highly unstable liquidity-
preference and investment-demand schedules are also clearly the product of an
unstable and consequently uncertain world. 4 Problems of uncertainty and specula-
tion in the foreign exchange market, together with unemployment and inflation, are
also crucial causes of some of the worst balance-of -payments problems.

1 See, for example, the classic analysis of a generalised Keynesian system by

Don Patinkin: Money, Interest, and Prices, Evanston, Illinois, 1956, Ch. XIV.
* J.M.Keynes: The Ueneral Theory ot Employment, interest, ana Money,
London, 1936, pp. 168-170.
* öee, ior example, me excellent analysis oy j.u.iooin: liquidity- irreierence
as Behaviour towards Risk, in: Review of Economic Studies, 1957-58.
* See, for example, Keynes celebrated discussion ot tne state ot long-term
expectation" (op. cit., Ch. 12).

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 17

Due to non-appropriability prob

instability is in general impossible
case for economic stabilising fisca
change-rate policies. Effective eco
a period may reduce markedly th
mitigating some of the most imp
expectations and unstable demand

4.4 Economic Growth

The classic problem of imperfect information is, however, that of eco-

nomic growth.
4.41 Uncertainty and Economic Growth. - Nowhere does uncertainty arise in
such an acute and obvious form as in choices concerning the future ; and, in
the case of growth, intertemporal choices are the very essence of the problem.
The basic choices in a growth context are those between present and future
consumption and between alternative types of investment including invest-
ment in research and in the development of new products and processes. In
making these savings and investment choices the private economic unit
(household or firm) must, in addition to knowing current alternatives, attempt
to forecast the future pattern of demand and supply in the relevant markets,
which will in turn depend upon the choices made by other private economic
units. The availability of refined market-research forecasts, the possibilities
of risk-reduction through diversification, the existence of well-developed
futures markets in money and certain basic commodities and the possibility
of long-term contracts all help to reduce the intrinsic uncertainties. In most
markets, however, even the best estimates are inevitably subject to extrem-
ely wide margins of error, and the scope for private action to reduce or avoid
these uncertainties is usually strictly limited. In a world of risk-averters it is
evident that savings, investment and growth rates would tend to be relatively
too low.
Some of the most serious privately-unavoidable risks facing the firm or
household in its savings and investment choices are those represented by
recessions and inflations. As we have seen in the preceding section, govern-
ment stabilisation policies can contribute greatly to the reduction of these
risks. Even with effective stabilisation measures guaranteed, however, the
privately unavoidable uncertainties involved in savings and investment de-
cisions would still be immense, and these uncertainties would particularly
affect decisions of crucial importance for growth, such as investment in re-
search and in the development of new products and processes.
Since these privately-unavoidable risks may vastly exceed socially-un-
avoidable risks, specific growth-oriented policies may be justified. Thus, for
example, private uncertainty can be further reduced by the techniques of so-
called "indicative planning", essentially a vast socialized market-research
and public-relations project. In particularly important fields such as research

1 See J.G.Head: Public Goods and Public Policy (loc. cit.), Section 4. 222.

2 Finanzarchiv N. F. 25 Heft 1

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 18 John 0. Head

and development, socialization

sidies may make possible a mu
risks and the achievement of
cies familiar to the public fin
and loss-offset provisions, als
reducing effects. Loss-offset
can be regarded as a special t
broaden the coverage and in
arrangements such as futures m
the stock exchange, may also b
cations of socialism is similarly
ved from the socialization of in
4.42 Irrationality and Economi
famous argument regarding t
sumption provides one of the
which has been directed at t
finance literature. This argum
certainty, of an irrational pre
pleasures has its immediate ap
a case for government promot
generally. Amongst the policie
ment would be compulsory sa
social insurance.
The irrationality case for growth-promotion is further supported by the
equally well-known intergeneration equity argument that time-preference is
not only irrational but immoral. According to this argument, the preferences
of present and future generations should on ethical grounds be weighted
equally in determining the allocation of resources over time. Like most
equity discussions to be found in the public finance literature, this intergener-
ation equity argument makes the same sort of appeal to the moral authority
of elites as the Pigovian irrationality argument.2 The two arguments thus
illustrate very well the importance of the authoritarian tradition in the
growth literature.
In spite of their undoubtedly controversial nature, they constitute a critique
of the individualistic tradition in a growth context, which has considerable intuitive
appeal to many "informed groups" and cannot safely be ignored in any comprehen-
sive welfare framework.

1 Vide supra, Section 3.122.

2 Indeed the intergeneration equity argument is often not clearly distinguished
from the Pigovian irrationality argument. This is true, for example, of Marglin's
otherwise excellent discussion referred to in footnote 1, p. 22 below. Pigou himself does
not in fact employ this further argument. His well-known contention (The Econo-
mics of Welfare, 4th edition, p. 29) that the government must act as "trustee for
unborn generations" in the field of conservation policy is based on an argument
regarding capital market imperfections (and hence uncertainty), not intergeneration
equity. See ibid., Part I, Ch. II, Section 4.
The irrationality and intergeneration equity arguments have also led to some
confusion regarding the "true meaning" of consumer sovereignty in an intertemporal
context. According to some writers, consumer sovereignty requires that the prefer-

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 19

The importance of such consider

more recent public finance literatu
an important role has been assign
count" based at least in part on th
ations. *

5. Information Problems vers

We have seen that individual bene

distortion problems constitute the
concept. Generalising this essentia
perfect information, we have fou
extremely important role in more
Keynesian theories of market failu
tion of resources, economic instab
growth are all to some extent ascr
rationality. Thus, in a generalised
Keynesian multiple theory of the p
stability, and growth, all exhibit imp
In recent discussions it has beco
can also be generalised into a com
Sidgwick's concept of non-appropr
economies.2

The analysis of Section 4 above shows that non-appropriability problems such

as those of external economies and decreasing costs play an extremely important
role in market failure of uncertainty origin. In particular, market failure problems
due to uncertainty can only arise where, for maximum efficiency, the risk-reducing
or insurance service must be provided in partially non-appropriable form and in-
volves large numbers of economic units. There is thus an extremely important con-
nection between the two theories obtainable by generalising the public goods and
merit goods concepts.
The two approaches are, however, by no means identical. Although there
are no Parefo-relevant uncertainty problems which do not involve non-ap-
propriability, there are other imperfect knowledge concepts, notably that
of irrational preferences, where some attempt is made to abstract completely
from non-appropriability problems ; and there are non-appropriability prob-
lems in which uncertainty or irrational preferences do not seem to be signi-
ficant. It may therefore be interesting to consider at least in general terms the
relative importance of uncertainty and irrationality problems compared with

enees of present individuals should be decisive. Others maintain that the preferences
of all consumers, future as well as present, must be fully represented. It seems simpler
and less confusing to define consumer sovereignty in terms of the preferences of
present individuals, and to recognise its possible welfare defects in terms of ration-
ality and intergeneration equity.
1 For an interesting recent discussion and further references, see M.S. Feld-
stein: The Social Time-Preference Discount Rate in Cost Benefit Analysis, in: Econo-
mic Journal, 1964.
2 See J.O.Head: Public Goods and Public Policy floe cit.J.

2*

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 20 John G. Head

pure non-appropriability proble

theory of the public household

5.1 Distribution

The main relatively pure non-appropriability problem in a distributional

context would appear to stem from the possibility that members of the
middle- and upper-income groups may derive altruistic satisfaction from in-
creased consumption by the lower-income groups. 1 A member of the upper-
income group unilaterally transferring income to the lower-income groups
may thus provide altruistic benefits to other high-income individuals. As a
result, private philanthropy will tend to be restricted because of non-appro-
priability problems, even though any given high-income individual might be
willing to contribute significantly more for the benefit of the low-income group
if other high-income individuals would agree to do likewise. If the number of
economic units involved is large, however, voluntary cooperation is likely to
be of only very limited adequacy. It may thus be possible to justify a sub-
stantial redistributive tax-transfer system for the benefit of the lower-income
groups imposing proportionate sacrifice on all members of the higher-income
groups.
Although equity has seldom been recognised as an important non-ap-
propriability problem, there can be little doubt that attitudes of this sort
contribute significantly to the relative general acceptability of redistributive
welfare schemes for depressed groups and of progressive income taxation
itself.
From our discussion of Section 4.11 above, it is clear that the uncertainty
problem is also likely to be a powerful if sometimes rather subtle force
affecting attitudes to distribution in a market economy. Although a clear
separation of motives in individual attitudes may be almost impossible to
achieve, it seems that uncertainty-related attitudes would usually be no less
important than purely altruistic or aesthetic egalitarian attitudes. To judge
from their importance in the literature, the various authoritarian approaches
to the equity question likewise cannot be neglected.

5.2 Resource Allocation

We have already cited cases such as a fireworks display, a bridge, or

theatrical performance, in which important elements of external economies
or joint supply in the public goods sense may exist without necessarily in-
volving any significant problems of individual benefit evaluation.2
Other examples of external economy or diseconomy problems frequently
analysed or quoted in the literature, such as the familiar case of apples and nectar,
or smoky chimneys, likewise involve no obvious problems of uncertainty or irration-

1 See the highly suggestive analysis by W.8. Vickrey: One Economist's View
of Philanthropy, in: F.Q.Dickinson (ed.): Philanthropy and Public Policy, New
York 1962, pp. 40-44.
* vide supra, section ó.ó.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 21

ality. Similarly, information problem

familiar examples of decreasing costs
due to elements of lumpiness or indi
indivisibility peculiar to the public g

As indicated in Section 4.21, the

bility problems in which informat
some of the best examples of public
al risk-reducing services, such as de
where the risk is not narrowly asso
modities. In some of these cases, su
extends to the evaluation of the benefits from the service itself. We have also
found many examples of uncertainties more narrowly associated with the
purchase of particular commodities, providing a possible justification for
pure foods acts, and other standards and safety legislation. Extremely signi-
ficant examples of conventional decreasing costs associated with uncertainty
are to be found in more formal insurance arrangements in such fields as hos-
pital and medical care, as well as in the case of many of the preventive ser-
vices already mentioned.
It is therefore clear that there are important examples of both pure and
uncertainty-related non-appropriability problems in the field of allocation.
Many alleged examples of irrational preferences are also to be found in the
literature, though their existence and importance are obviously more difficult
to demonstrate.

5.3 Economic Stability

In a stability context, the most important example of a non-appropria-

bility problem in which uncertainty does not seem to be an essential factor
is provided by cost-inflation.1 Even under relatively stable monetary and
demand conditions, employer or employee organisations may attempt to use
their market power over price to alter the distribution of factor income in
their own favour. Other groups in a position to pass on such a redistributive
price-increase and thus protect their factor share proceed to do so, and a
cost-inflationary spiral ensues. Losing groups usually include farmers, pen-
sioners, salary-earners, and others whose incomes tend to be relatively fixed
in money terms. 2 Under these conditions, the full social benefits to be derived
from price-, dividend-, and wage-restraint will not in general be appropriable
by the individual union or firm, and public policy (notably an incomes policy)
must be relied upon.
As shown in Section 4.3, uncertainty plays a crucial role in modern
theories of economic instability of demand-side origin.

1 For a very useful recent survey of cost-inflation theory, see M . Bronfenbrenner

and F. D. Holzman: Survey of Inflation Theory, in: American Economic Review, 1963,
Section III.
* It is not tnereîore the case that only ignorance ot the ability ot other groups
to defend their income shares can explain the original wage-push or profit-push.
Although this and other imperfect knowledge problems, such as money illusion, are
no doubt frequently of some importance in cost-inflation, they do not seem to be an
essential ingredient.

This content downloaded from

106.120.213.173 onffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 22 John G. Head

Thus we saw that unstable demand schedules, interest-elastic demand for

money, interest-inelastic investment demand and elastic expectations are all phe-
nomena of an uncertain world. Although interest in the problem of cost-inflation
has been steadily increasing over the post-war period, it still seems reasonable to
suggest, in the light of the available evidence, that uncertainty-related non-appro-
priability problems are of overwhelming importance in a stability context.

5.4 Economic Growth

The pure non-appropriability problem which has been most fully ana-
lysed in a growth context is that in which the consumption of future gener-
ations enters altruistically into the utility functions of members of the present
generation. 1 Activities of one member of the present generation which lead to
increased consumption by future generations provide altruistic benefits to
other members of the present generation. Growth-promoting activities there-
fore tend to be underexpanded because of non-appropriability problems.
Under these conditions, an individual for whom the sacrifice involved in
additional saving would not be fully justified in terms of greater consumption
by future generations may nevertheless be willing to save more if other mem-
bers of the present generation could be relied upon to do likewise. Since a
large number of private economic units will usually be involved, voluntary
cooperative solutions are likely to be seriously inadequate, and growth-pro-
moting government policies apportioning the sacrifice of present consump-
tion may therefore be justified.2
This sort of "future-oriented altruism" is probably relatively unimportant in
highly-developed Western societies. It may, however, be a socio-political factor of
some importance in certain underdeveloped countries.
More important in most cases, no doubt, is the phenomenon of "future-
oriented self-interest" . Since the benefits in terms of higher real wages and con-
sumption levels resulting from growth-promoting activities spill over onto
other individuals, members of the present generation wishing to enjoy higher
levels of real income and consumption in the future may be willing to bear an
increased burden of saving, risk-bearing, and dividend- and wage-restraint
only if others are forced to do likewise. Thus it is future-oriented self-interest,
i.e. the promise of more rapid rates of growth and thus higher living standards
for the present generation within a reasonably short planning period, rather
than future-oriented altruism, which appears to account for a significant part
of the genuine support for comprehensive growth policies in many countries.

1 See, for example, the comprehensive analysis by S. A.Marglin: The Social

Rate of Discount and the Optimal Rate of Investment, Quarterly Journal of Eco-
nomics, 1963.
A When simultaneous account is taken ot luture-onented altruism and the
present-oriented altruism discussed in Section 5.1 above, the marke
suboptimal growth rates due to future-oriented altruism may be offset
to super-optimal growth due to present-oriented altruism. In par
present-oriented altruism, individual members of the middle- and
groups will tend to want fellow-members to transfer part of their
income individuals in the present generation. For some discussion of th
points, see the "Comments" on Margliris article by G.Tullock,
D. Usher, Quarterly Journal of Economics, 1964.

This content downloaded from

106.120.213.17fff:ffff:ffff on Thu, 01 Jan 1976 12:34:56 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 23

As we have seen in Section 4.4, h

overwhelming importance in a gro
significant degree of "future-orien
terest" may be expected, notably
tainties and risks associated wit
frequently the greatest. In general
related non-appropriability proble
non-appropriability problems in a
sion1, it is clear that irrationality
able importance.

6. Merit Goods and Welfare Politics

We have now examined the role of generalised merit goods (i.e. know-
ledge) problems in the economic theory of the public household and com-
pared them with the generalised public goods (i.e. non-appropriability)
problems which have figured so prominently in recent discussions.
The great revival of interest in public goods as a crucial concept in the
theory of market failure has, however, also been accompanied by a revival
of interest in the political problem of improving on market performance. In
particular, it has been increasingly recognised that, where the market fails,
it does not necessarily follow that the political mechanism can reasonably
be expected to do better. The economic theory of the public household, based
on the theory of market failure, requires therefore to be supplemented by a
political theory of the public household. 2

6.1 Information Problems and Welfare Politics

It appears that there is in fact a very close relationship between the
central inefficiency concepts in the generalised merit goods theory of market
failure and those in the theory of political failure. Thus the theory of welfare
politics suggests two major reasons for the failure of the political mechanism:
the government may not try to achieve the welfare objectives postulated in
the economic theory of the public household ; and even if it tries, it may lack
the knowledge necessary for even approximate success. These problems clear-
ly correspond to the irrationality and uncertainty problems of the generalised
merit goods theory of market failure.
1 Vide supra, Section 4.42.
2 Important contributions to the development of such a theory have recently
been made by Downs and Buchanan and Tullock. See A, Downs: An Economic
Theory of Democracy, New York 1957; and J.M.Buchanan and G. Tullock: The
Calculus of Consent, Ann Arbor 1962. Increasing attention has also been paid to the
pioneering work of Wicksell and Lindahl. See, for example, J.G.Head: Lindahl's
Theory of the Budget, Finanzarchiv, N. F. Bd. 23, 1963-64. For a general survey of
the contributions of Wicksell, Lindahl, Downs, and Buchanan and Tullock, see
J.G.Head: The Welfare Foundations of Public Finance Theory (loc. cit.), Section 5.
On Lindahl and Downs, see also E. Liefmann-Keil : Zur Entwicklung der Theorie der
Bewilligung öffentlicher Einnahmen und Ausgaben, in: Finanzarchiv, N.F. Bd. 19,
1958-59.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 24 John G. Head

6.11 Irrationality. - The prob

illustrated by the case of abso
such cases it would obviously
tives happened to correspond
there will usually be only the
typical absolutist objectives
prestige, the establishment of
such welfare objectives as equi
course, partly explains the scep
al in conservative and liberal c
per se as a panacea for market
torically.
This is not, of course, to suggest that no political mechanism can reason-
ably be expected to improve on market performance unless the government
is directly inspired by an interest in the common good.
Thus in Downs* theory of the democratic political mechanism 2, politicians are
assumed to be motivated only by the desire to be elected. Competition between poli-
tical parties tends, however, to force the adoption of Parefo-optimal policy program-
mes in order to maximise electoral prospects. 8 It was in fact a familiar argument in
the public finance literature of the late 19 th century (e.g. Sax, De Viti de Marco,
and Mazzola) that some approach to the achievement of a Parefo-optimum is a
necessary condition for political equilibrium in a democracy. 4

6.12 Uncertainty. - The problem of uncertainty is obviously no less important.

Thus, even where the government is directly or indirectly motivated by a
concern for welfare objectives, government uncertainty regarding given in-
dividual preferences, the nature of existing inefficiencies and the effects of
alternative policies, together with individual uncertainty regarding the
nature and effects of government activities, may seriously weaken the link
between political aims and achievements.
Downs, for example, has emphasized that it is only on the completely
unrealistic assumption of perfect knowledge that his majority voting model
can conceivably be expected to produce Parefo-optimal economic policies.
In the more general case, uncertainties of all sorts will prevent the system
from achieving any more than the roughest approach to Parefo-optimality.
Under these circumstances, Parefo-efficiency will no longer be a necessary
condition for maximisation of electoral prospects. 5 Within fairly wide limits,
government programmes can thus come to reflect the private motives and
prejudices of political parties and the empire-building tendencies of the bu-

1 As Scitovsky has pointedly remarked (op. cit., p. 242): "For many centuries,
food, clothing, and shelter symbolized market goods, while cathedrals, palaces, and
armies were the symbols of collective spending".
2 Downs, op. cit.
8 For a careful analysis of this tendency, see ibid., pp. 178-182.
4 De Viti 8 discussion (op. cit., Book I, Ch. I, Section 7) is perhaps the most
suggestive.
6 For Downs' analysis of the effects of uncertainty, see op. cit., Ch. 10, Sections
IIC-IIF. Also "Why the Government Budget is Too Small in a Democracy", World
Politics, 1959-60.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 25

reaucracy ; and the electoral proble

programmes to an uncertain electo
advertising, emphasizing simple a
all sorts of irrelevant emotions.

6.2 Merit Goods and the Political Mechanism

In spite of these general limitations of the political process, is there per-

haps some reason for assuming that the political mechanism is uniquely
capable of improving on market performance in the case of merit goods ?
Curiously enough, some such assumption does seem to be implicit in
much of the traditional public finance literature. As pointed out in Section 3.3
above, individual benefit-evaluation difficulties have frequently been held to
constitute the crucial problem posed by the existence of public goods. Thus
it is argued that the concept of individual benefit shares is not applicable in
the case of public goods where the benefits accrue collectively, i.e. to society
as a whole, rather than to individuals.1 Furthermore, pointing to examples
such as defence, education, growth, and foreign aid, it is suggested that the
benefits are anyway too indirect, long-run or technically complicated to be
capable of accurate individual assessment. 2 Reliance upon the expert know-
ledge of the bureaucracy or government is therefore essential; or, in more
democratic versions of the argument, the special value of collective decision-
making in these cases is vigorously asserted.
In fact, however, there are very good reasons for supposing that the ir-
rationality and uncertainty characteristics of merit goods pose far more dif-
ficult problems for the political mechanism than those associated with pure
non-appropriability.
6.21 Irrationality. - The political problem in cases of irrationality can readily
be illustrated by considering the Pigovian example of an irrational preference
for present over future consumption. Abstracting from all other knowledge
problems, could, for example, a majority voting mechanism of the Downsian
sort be expected to lead to the implementation of the growth policies required
to correct the resulting inefficiency ? Since the vote-maximising political
parties of the model have no interest in the achievement of welfare goals for
their own sake but only for their vote-getting potential, the answer depends
on whether additional votes could be attracted by promoting a rate of growth
in excess ofthat consistent with the true intertemporal preferences of electors.
It is clear, however, that each and every individual would prefer a given
Parefo-optimal programme excluding Pigovian growth measures to the same
programme plus (non-redistributive) Pigovian growth measures. The adop-
tion of the Pigovian policies would therefore result in electoral catastrophe

1 As we have already seen, this argument was effectively demolished by De

Viti. See footnote 2, p. 9 above.
¿ uart ofioup, lor example, particularly emphasizes tñe benefit-evaluation
problems arising from the preventive nature of so many important government ser-
vices.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 26 John G. Head

and could not reasonably be ex

of the model. *
The argument for reliance up
of government for the perfor
therefore politically extremely
to rely upon the emergence o
truistic desire to rectify irrat
mocratic process, as envisaged
performing this function, it
theory is completely non-oper
Following an important line
literature, it might, however,
assumes complete identity of i
political process. According to
social psychological forces (
Sax)2, individuals tend to tak
views in the context of collect
than in the market. 3 Thus it
that individuals take a complet
eign aid in a political context f
ket behaviour.4 The vote-max
mocracy would therefore be fo
merit want functions.

1 As an example of an irrational preference, the Pigovian case has the interest-

ing peculiarity that an electorate which ex ante may unanimously oppose the requi-
red growth measures, may equally approve them ex post, if they can somehow be
forced through initially. Indeed, even more Draconian measures would seem justi-
fied ex post on Pigovian arguments, since, as Pigou himself emphasizes (op. cit.,
p. 117), the defective telescopic faculty also works in reverse (when we contemplate
the past). The fact remains, however, that no government could ever be elected in a
Downsian democracy on a programme of Pigovian growth measures, or even on the
continuation of such measures after a period of growth-oriented and ex post justified
dictatorship.
2 See Emit ¡Sax: Die Wertungstheorie der bteuer, Zeitschrift tur Volkswirtschaft
und Sozialpolitik, 1924. Most of the relevant section is available in an English trans-
lation and summary: The Valuation Theory of Taxation, in: R.A.Musgrave and
A. T. Peacock (eds.): Classics in the Theory of Public Finance, London and New York
1958.
3 From the public finance literature, in addition to the important work of Sax,
see, for example, E.R.A.Seligman, The Social Theory of Fiscal Science, Political
Science Quarterly, 1926. A similar view has been persuasively advocated over a long
period by Gerhard Colm. See his Volkswirtschaftliche Theorie der Staatsausgaben,
Tübingen 1927; Essays in Public Finance and Fiscal Policy, New York 1955; Com-
ments on Samuelson's Theory of Public Finance, Review of Economics and Statistics,
1956; In Defense of the Public Interest, Social Research, 1960; and, most recently,
in his article "National Goals Analysis and Marginal Utility Economics", Finanz-
archiv, N.F. Bd. 24, 1965. In his latest contribution (op. cit., p. 213), Colm sum-
marizes his position as follows: "As there is no strict comparability between the pre-
ference scales of two individuals, there is also no cogent relationship between the
preference scales of the same individual as homo oeconomicus and homo politicus."
4 The possibility of applying this argument m the held of growth is considered
(and rejected) by Marglin (op. cit., pp. 98-99).

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 27

Although the existence of such d

argument scarcely suggests the ex
red in the present context. A mor
essentially cooperative social psych
lectivism" merely help to overcom
large numbers which prevent the
tion of given individual preference
a more accurate expression of giv
way suggests that it can be used to
more obvious explanation of the u
that individuals are extremely unc
from changes in the output of pub
therefore highly unstable and can
the differing influences of the mar
6.22 Uncertainty. - We have alre
regarding given individual prefere
and the effects of alternative polic
regarding the nature and effects o
a rough approach to a Pareto- optim
ical mechanism. If, due to incomp
know his own true preferences for
to an already formidable list. Is th
vidual's true preferences for thes
expression through, for example,
visaged by Downs than through th
upon whether more adequate provi
the prospect of additional votes. In
cases tend to be indirect, delayed
appreciate1, and given a general av
other things being equal, a Parefo-
the relevant merit good policies w
programme providing benefits wh
even if government experts or par
preferences of individuals, the incl
electoral disaster. Such policies may

1 Uncertainty regarding the benefits

uncertainty regarding the benefits of
insurance service. In fact, however, un
lem in the case of many of the mo
those of a preventive character.
2 Appropriate merit goods policies w
ate as fully justified ex post, if they c
in contrast to the Pigovian merit go
(see especially footnote 1, p. 26 above),
of continuation of existing merit goo
sures could not safely be omitted from
mise its electoral prospects. It may th
successfully to introduce appropriate

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 28 John G. Head

Writers in the democratic po

that the basic Downsian model
sion and deliberation under th
elites. According to this argum
mechanism is that individual p
course of the free discussion which is held to be essence of the democratic
process. x It is, however, possible to doubt whether the competitive political
advertising of the democratic process is any less misleading or any more con-
ducive to individual self-knowledge than advertising in the market.

7. Conclusions

The main results of the above analysis can be summarized as follows:

(a) Musgrave's concept of a merit good appears to exhibit as many as
three distinct and independent characteristics: individuals find its benefits
difficult to evaluate; it is frequently the subject of special distributional
objectives; and it typically exhibits public good aspects. Although there is
no necessary relationship between these characteristics, the merit goods con-
cept is extremely useful in focussing attention on the empirically very signi-
ficant coincidence that some of the best examples of goods posing individual
benefit-evaluation difficulties are also the subject of special distributional ob-
jectives and exhibit important public goods characteristics. The concept thus
provides a particularly striking and important illustration of compound
market failure.
(b) Individual benefit-evaluation difficulties are clearly intended by
Musgrave to constitute the essence of the merit goods problem. Generalising
this essential characteristic to cover all types of imperfect knowledge, it is easy
to show that the information problem also plays an extremely important role
in more familiar Pigovian, Keynesian and post-Keynesian theories of market
failure; maldistribution of income, misallocation of resources, economic in-
stability, and non-optimal rates of economic growth are all to some extent
ascribable to problems of uncertainty and irrationality. Thus, in a more
general sense, the familiar objectives of the post-Keynesian multiple theory
of the public household, namely equity, efficiency, stability, and growth, all
therefore exhibit important merit goods characteristics.
(c) In recent discussions it has become clear that the concept of a public
good can also be generalised into a comprehensive theory of market failure
via SidgwicFs concept of non-appropriability or BaumoVs generalised external
economies. Our analysis indicates that generalised public goods problems also
play a crucial role in the generalised merit goods theory of market failure. In
particular, market failure due to uncertainty can only occur where, for max-
imum efficiency, risk-prevention or insurance services must be provided in

but electorally irrelevant unpopularity. There will, however, be little or no political

pressure to do so in the basic Downsian model.
1 From the public finance literature, see, for example, De Viti de Marco, op. cit.,
p. 118, footnote 2.

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms
 On Merit Goods 29

partially non-appropriable form. T

nection between the two theories o
the public goods and merit goods
(d) The two theories are, however
are no Pareto-velevsnit uncertaint
propriability, there are other imp
irrational preferences, where an at
non-appropriability problems ; and
which uncertainty or irrational pr
therefore seems desirable to prese
types of problem. An examination
and irrationality problems compar
clearly suggests that a general econ
afford to emphasize either set of p
(e) It is interesting to observe th
lems of market failure and those
the failure of the political mechan
to achieve the objectives postulate
household, and that, even if it trie
even approximate success. These pr
ality and uncertainty problems of th
failure. On examination, it appear
on market performance is likely to
goods than in the case of pure pub

This content downloaded from

106.120.213.173 on Wed, 04 Nov 2020 08:17:40 UTC
All use subject to https://about.jstor.org/terms