View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

George J. Stigler, professor of economics, Columbia University
The members of the legislative, executive, and judiciary de­
partments of 13 and more States, the justices of peace, officers
of militia, ministerial officers of justice, with all the county,
corporation, and town officers, for 3 million and more of peo­
ple, intermixed, and having particular acquaintance with
every class and circle of people, must exceed, beyond all pro­
portion, both in number and influence, those of every descrip­
tion who will be employed in the administration of the Fed­
eral system (The Federalist, No. 45).
The preservation of a large role in governmental activity for local
governments is widely accepted as an important social goal. No one
can doubt that the individual citizen gains greatly in political dig­
nity and wisdom if he can participate m the political process beyond
casting a vote periodically. I t is also generally conceded that a good
political system adapts itself to the differing circumstances and mores
of different localities, or, as I would wish to rephrase it, the system
should allow legitimate variations of types and scales of governmental
activity to correspond with variations m the preferences of different
groups of citizens.
Nor will it be denied that this social goal is being increasingly
sacrificed. In 1900, virtually all questions of housing, public health,
crime, and local transportation were dealt with exclusively by State
or local governments, and the role of the Federal Government in edu­
cation, regulation of business practices, control of natural resources,
and redistribution of income was negligible. Today the Federal Gov­
ernment is very active in each of these areas, and its share of responsi­
bility is gradually increasing.
I propose to examine some of the reasons which are given for the
growing centralization of political processes. The proper range of
activities of government in general will not be raised. Our question
is simply th is : I f the people in a given community wish to embark on
a particular governmental policy, when does the efficient discharge
of this policy require that it be imposed by a central authority also
upon other communities ?
In many minor areas of governmental activity no real questions are
raised, as yet, about the feasibility of local sovereignty. I f a given
community wishes to have superb library facilities, it can build and
pay for them ; if another community wishes instead a skating rink, it
may so choose. I f individual citizens in any community disagree
strongly with the majority preferences, they may move to a more con­
genial community. Since governmental functions must often be pro­
vided upon a considerable scale to be tolerably efficient in execution, a




sufficiently eccentric individual may not be able to find any community
with enough like-minded individuals to be able to adopt th at series
of governmental policies which would exactly suit his taste. F o r exam­
ple, he may wish to live in a community with gravel streets and a
magnificent observatory, and find no community willing to provide
this combination. This sort of limitation is also encountered in con­
suming the products of private enterprise—I may not find precisely
the automobile or typewriter that suits me.
In most areas of governmental activity, however, it is increasingly
felt that local governments are inefficient units. When any of three
types of governmental activity are sought, it is said that the unit of
effective administration must be large in scale—
1. When the object of a regulatory policy can be nullified by
the competition of (including migration to) other local govern­
mental units.
2. When the source of revenue of the activity can escape finan­
cial responsibility by migration to another unit.
3. When the policy is incapable of efficient performance upon a
local scale.
We consider these problems in turn.

h e





o m p e t it io n

Suppose that a community wishes to set a high standard of factory
safety, and requires the installation of a very expensive safety device.
Then the local portion of a much larger industry will be undersold in
the common market by factories in other communities, provided they
do not also simultaneously set as high standards of safety practices.
The local branch of the industry then dies or migrates. These facts
can be taken as data for our discussion.
The essence of this argument is th at competition, which usually
works so well in the area of private enterprise, serves to defeat desir­
able goals in the area of government. I f every governmental unit, save
one, were to desire and require elaborate safety devices in the facto­
ries of some industry, it is claimed that their desire could be stultified
by the presence of the exceptional community which did not have this
desire, because the regulated industry would migrate to this commu­
nity and escape regulation, and the knowledge that it would do so is
often enough to prevent the various communities from attempting to
regulate it.
I t may be remarked that a similar argument is often encountered
in the private-enterprise sector. Plants with low wage rates, it is
often said, force plants with high wage rates to reduce their wages in
order to compete successfully in the common market. In this case the
argument is reversible: the plants with high wage rates force plants
with low wage rates to raise their wages in order to compete success­
fully for workers in the common labor market. Both formulations,
however, are singularly uninformative, for they do not lead directly
to the correct conclusion, which is that the wages of all (similar)
workers must approach equality in all plants under competition, and
the common wage rate will be governed by the value of the worker’s
services in those plants which can pay this rate. Can it be th at some
parallel obscurity attaches to the customary formulation of the unfor­
tunate effect of competition among governments ?



The governmental analysis is, in fact, incomplete. Suppose any
community set the required level of safety practices as high as it
wishes, and then gave a subsidy to each enterprise in the locality equal
to the additional cost that these safety devices imposed upon the en­
terprises. Then there would be no tendency for the local industry to
be handicapped in competition with other areas with lower safety
standards, and the community would enjoy more worker safety and
less of other things than other communities. I f 47 percent of the lo­
calities or 99 percent of the localities embark upon this policy, then 47
or 99 percent of the factories will have the desired safety practices,
and the nonconformist competitors will not have the slightest ten­
dency to injure or attract these safe and expensive factories.
When a community imposes the safety regulations without giving
a compensating subsidy, its troubles arise from the fact that it is seek­
ing to push these higher costs off on consumers, and neither local nor
distant consumers wish to assume this burden. The problem of com­
petition resolves itself into an unwillingness of the community to bear
the costs of its policy when they are posed as an explicit burden.
A similar analysis holds when the community wishes to require of
some consumer good that it be of unusually high quality. I f it spec­
ifies that only goods of this unusual quality be sold in the community,
the producers will be quite eager to meet the specifications—at a re­
munerative price.
Although it involves a digression, it may be profitable to discuss
more generally our example of factory safety devices because the dis­
cussion wTill serve to illuminate the workings of competition in general.
If workers are faced with the choice of working in one plant, un­
equipped with safety devices, at an hourly rate of $1.50 but wTith ex­
pected losses from injuries of 5 cents per hour, and in another plant
with safety devices they are offered $1.46 w’ith no expected losses
from injuries, we should expect them to choose the latter plant. I f
under these conditions they do not choose the safer plant, the most
probable explanation is that they do not correctly appraise the ex­
pected losses from injuries and the remedy is to inform them of the
consequences of working in factories unequipped with safety devices.
In a fully competitive system the entrepreneurs will supply at cost
all the safety devices that the workers demand, and all safety devices
which return (to the worker, in terms of reduced injuries) as much
or more than the cost wil ] be adopted. I t may well be th at in this
situation there will be safety devices wThich do not pay but which
would reduce injuries further, and that the community as a whole
sets a higher value on avoidance of these injuries than the workers
themselves do. Some moral philosophers might argue that these work­
ers should set a higher value on the avoidance of injuries, but the
workers do not, and in a society with free choice of occupation they
cannot be made to pay for more safety than they wish. Hence the
society must bear the costs of achieving more safety, and the sole
question is whether the costs be borne by consumers through compul­
sory installation of the safety devices and restriction of supply, or by
direct grant from public funds.
The competition of other communities as tax collectors is an im­
portant form of the alleged difficulty arising out of competition. Sup­
pose community A wishes to have splendid and expensive schools,



streets, housing, poor relief, and what not. I f it levies sufficient taxes
to finance this elaborate program, a large portion of the tax base (in­
dustries and well-to-do individuals) will leave the community w iile
simultaneously a large number of beneficiaries of the generous pro­
gram may immigrate. The tax rates on the narrower tax base will
have to be prohibitive (from the viewpoint of the remaining taxpay­
ers) to finance the sumptuous program.
Again we can accept the facts, with one temporary amendment.
Let us assume that the same income is received by every family, and
no questions of income redistribution are involved. W ill the presence
of communities with lower tax rates defeat the ambitions of com­
munity A ? The answer is clearly in the negative. There will be some
redistribution of population among communities: those people who
prefer cheaper public services and lower tax rates will move else­
where, and others with opposite taste will move to A. Competition of
communities offers not obstacles but opportunities to various com­
munities to choose the types and scales of governmental functions
they wish. The proviso that all family incomes are equal has a vast
influence on this argument, of course, and we turn now to income
redistribution as a goal.

h e


is t r ib u t io n



ncom e

I f all families had equal (real) income, would there be any need
for local governmental units ? Why could not each city be a private
corporation, supplying at a price the services its dwellers demanded ?
W ith many, many such corporations, competition would prevent mo­
nopolistic pricing, and schooling and police and fire protection would
be sold at a price including a fair rate of return on investment. This
scheme would obviously be inappropriate where the service must be
a monopoly (like national defense) and probably also where the com­
munity size was so large (due to the economic advantages of size) that
the communities were too few to rely upon competition, but let us put
these instances aside. We are not seeking to prove that there should
be no government, but rather to find the logic of government at the
multiunit governmental level.
A basic deficiency in this private enterprise organization of social
life, we would all agree, is that it allows excessive freedom to the indi­
vidual. I t would allow parents to horsewhip children, and it would
create communities populated chiefly with drunkards and drug ad­
dicts—although thieves would presumably prefer to live among honest
men (even with their policemen) than only with other thieves. Pub­
lic opinion would curb many undesirable personal actions, but the
society would wish to compel observance of its basic values. As a
result, we must recognize the need for political units large enough so
their numbers include enough normal people to insure the imposition
of the society’s basic moral standards on local communities. Our
States—with 1 or 2 possible western exceptions—meet this condition of
statistical large numbers.
The second basic weakness—some will call it a strength—of the p ri­
vate enterprise organization of local government is th at it would not
permit price discrimination; it does not have the ability to redistribute
income. The purely competitive organization of local services would



make it impossble for a local government to obtain money from the
rich to pay for the education of the children of the poor, except to the
extent that the rich voluntarily assumed this burden.
How can local governments cope with this problem? I f 99 com­
munities tax the rich to aid the poor, the rich may congregate in the
hundredth community, so this uncooperative community sets the tune.
Here competition does not perform with its usual excellence, for com­
petition is the system calculated to organize only voluntary activity.
W hat is the correct amount of redistribution of income in light of
the society’s desires ? I t is more than the unrestricted competition of
tax-free colonies of the rich would allow, but less than the most ag­
gressively egalitarian community would desire. The decision must
be in some sense a national decision, for the proper amount of redis­
tribution, even if rich and poor were chained to their communities,
could not depend upon the accidents of income composition of a par­
ticular community. And once this level of redistribution is set, no
one community may complain if its rich citizens migrate when it seeks
to go above this level of distribution unless the society is prepared to
let the most egalitarian community set the scale of income redistri­
Since redistribution is intrinsically a national policy, it should not
be restricted to a community level; a community consisting only of
poor people should receive the desired minimum social services.
Hence, in pure principle, the Federal Government should collect the
progressive levies and redistribute them (in whole or in part) to local
units with each unit receiving an amount governed by the number of
its poor and the degree of their poverty.
Given this system of tax revenue redistribution, the local govern­
ments could still be allowed to perform any function which they were
competent to perform efficiently. One community might choose to
spend more on schools and less on hospitals than another, but this is
surely an area of legitimate freedom; there is no “correct” distribution
of expenditures among such functions.
In a society which has no serious program of income redistribution
(even as a means to the attainment of minimum goals), local govern­
ments would face no basic revenue problems because of competition.1
I t is in keeping with this argument that a century ago almost all func­
tions were local and the problem of competition for the tax base was
negligible. W ith an appropriate fiscal system we could restore these
revenue considerations to a position of unimportance even in an era of
extensive income redistribution. There still remains the question of
whether the local governments could efficiently perform the enlarged
range of functions th at modern governments have assumed. We
turn now to this question.

h e


c o n o m ie s




H o w large must a governmental unit be to perform efficiently the
activities which the public wishes governments to perform? This is
an area which deserves much more attention than it appears to have
received, and the following remarks are highly tentative.
1 P e r h a p s a q u a lific a tio n s h o u ld b e e n te r e d w ith re s p e c t to th e g ro w th o f ta x a b le w e a lth
t h a t e sc a p e s a g e n e r a l p ro p e r ty ta x . I n E n g la n d th e d e s ire o f p ro p e rty o w n e rs to ease
t h e i r ta x b u rd e n s w a s a fo rc e in th e e m a s c u la tio n o f lo c a l g o v e r n m e n t; see E . C a n n a n ,
T h e H is to r y o f I^ocal R a te s in E n g la n d , sec o n d e d itio n , 1912, ch. V I.



. There are a set of functions which are intrinsically national because
they are indivisible. The greatest of these is national defense, and
it would be ill-served if each State or local unit were to undertake
the defense of its own area. One may cite also foreign relations, the
national governmental machinery, and the control of relationships
among lower governmental levels.
‘ In addition to such traditional functions, one may list certain func­
tions which are or can be performed at a local level but which must
be coordinated to achieve efficiency in their design. The transporta­
tion systems of localities must take some account also of the needs of
long-distance transportation. The radio and television stations of
various localities must not jam one another. These are functions
which in the economist’s language, have large external economies or
diseoconomies accruing to the areas which do not participate in their
execution, so it is essential th at they be formulated (although not
necessarily administered) on a larger area than the local government.
We should reserve for the Federal Government those functions
which are much more efficiently discharged on the largest scale.
When local performance involves large duplication, it is inefficient.
Thus it seems undesirable to have 48 estimates of wholesale prices
since the price movements in most regions will be parallel; on the
other hand, the calculation of cost-of-living indexes might suitably be
removed from the BLS to the States.
The optimum scale of performance has tacitly become identified
with the National, or at least the State, scale almost without examinmg the nature of the governmental functions under discussion. This
seems most surprising to the student of industrial organization; he is
accustomed to finding that the activity in an industry with a complex
technology is usually efficiently conducted by a firm smaller by almost
any measure than the government of a town of 25,000. Is there some
special characteristic of governmental functions that makes large
units necessary to efficiency ?
Only one characteristic seems a possible candidate for this role: the
great variety of functions performed by even the small governmental
units. The lack of specialization is pronounced even though politi­
cal scientists complain of a multiplicity of overlapping local units
(many of which were established to evade tax or debt limits on local
units). Some of these functions can be performed efficiently on a
very small scale. Many of the most distinguished private schools
and colleges are much smaller than the school system of a town of
5,000 people. Others are more varied. A police department can
efficiently control local traffic on a small scale; in one sense it must
be worldwide to have an efficient “missing persons” bureau.
But this variety of function is not really unusual. Every enter­
prise must use goods and services, or produce goods or services, which
must be produced or sold on a much wider scale than the enterprise
itself can undertake. Even a huge department store is not large
■enough to make its own delivery trucks, or to prin t the newspapers
in which it advertises. Ju st as cooperation in these matters is brought
about by the price system, so cooperation among governmental units
has been developed—and could be carried much further—to avoid the
determination and execution of all public functions by that govern­
mental unit which is most efficient in conducting the function with the
largest scale of operation.



I t happens, as we have already noticed, that one function of para­
mount importance must be conducted on a very large scale: the col­
lection of revenues designed to redistribute income. Much central­
ization, in fact probably most centralization, has been a consequence
of this situation. A central government is loathe to make grants
without exercising a degree of control over the local units which
disburse the funds. No degree of control less than 100 percent, how­
ever, is sufficient to guarantee local performance exactly as the cen­
tral authorities wish it, and there is no obstacle except tradition to
slow down their gradual extension of controls.
The case for imposing controls over the smaller units receiving
rants, however, is far from general. The central disbursing authority
as no monopoly of wisdom. The State boards of education have
imposed a series of certification requirements on local teachers, for
example, that have done much to lower the quality of elementary
education in the United States. When central governments have
superior civil servants, as they often do, the cause lies more often in
their control of finance and authority than in the advantages of
centralization. I t may be true that when most administrative units
are small the ablest men cannot conduct affairs on the largest scale,
but this seems an odd consideration to give weight in setting the func­
tions of local governments in a democracy. More often the complex­
ity of the tasks at the national level has reached such levels that not
the ablest men can control them efficiently.
If grants were given to local governments without supervision,
there would be some instances of gross neglect or venality and more
variety in the quality of the performance of public functions. We
should also expect to find that much of this variety was eminently
sensible, and that many types of experimentation would constantly
be embarked upon by the more venturesome and the more foolish
communities—with large social benefits from both the successes and
the failures.
I f we give each governmental activity to the smallest governmental
unit which can efficiently perform it, there will be a vast resurgence
and revitalization of local government in America. A vast reservoir
of ability and imagination can be found in the increasing leisure time
of the population, and both public functions and private citizens
would benefit from the increased participation of citizens in political
life. An eminent and powerful structure of local government is a
basic ingredient of a society which seeks to give to the individual the
fullest possible freedom and responsibility.