Macro Theoretical Approaches to Public

Policy Analysis: The Fiscal Crisis

of American Cities

By HENRY TEUNE

ABSTRACT: Most empirical research on the consequences

of public policy follow a mirco, "historical," and incremental
approach. Thus, the validity of the predictions have limited
generality. A macro, theoretical, and structural approach is
presented as an alternative. Such a reorientation to public
policy analysis will not only provide understanding of the
context of policy predictions, but also opportunities for test-
ing them both historically and comparatively. It would also
bring together social science theory and policy research.
This approach is illustrated with three macro structural
theories of the urban fiscal crisis in the U.S.—the conflict
between political demand for welfare and capital for economic
growth; the decline of the city as a political force in national
politics of economic distribution; and the changing techno-
logical base of cities and their economic viability. Each of
these theories provide for different interpretations and predic-
tions of policies that are likely to be effective in achieving
collective goals.

Henry Teune is Professor and Chairman of Political Science at the University of

Pennsylvania. He is co-author of The Logic of Comparative Social Inquiry, The
Integration of Political Communities, and Values and the Active Community. He is
participating in a cross-national study of local governments which focused on
leadership in a sample of United States cities and comparable local governments
in other countries. A similar study on change in local political governments is now
underway. He is currently completing, The Developmental Logic of Social Systems
and co-editing The Social Ecology of Change: From Equilibrium to Development,
both to be published in 1978.
174
 175

AMONG the problems of policy theory, which is both historically

analysis that is intended to (cross-time) and comparatively
increase knowledge for public (cross-system) based, that this paper
choice are: is directed.
1) the predictions of outcomes
have limited historical and cross- PARADIGMS OF PUBLIC POLICY
system validity; 2) the predictions do ANALYSIS
not discriminate between policy
constraints and policy instrumental- Contemporary public policy
ities ; and 3) the predictions are not analysis either is handicapped by or
formulated in a way that they can be is the beneficiary of several types of
tested outside of the specific context paradigms-public choice theory,
to which they are intended to be or quasi-experimental analysis, norma-
have been applied. tive evaluation, game theoretic
Public policy, in this discussion, is models.1 Whatever their relative
distinguished from policy in gen- merits, two empirical conditions for
eral. The latter can be conceived as the application of any paradigm of
the total set of goals that are both public policy analysis are necessary,
if the results of the analysis can be
explicitly and implicitly operative used for purpose, political purpose.
in a social system without regard to
their origin (for instance, public or First, the subjective states, the
purposes of the actors, must be dif-
private institutions), their normative
character (for instance, behavior in- ferent from, incongruent with, the
states of the system. If this is not
ducing or behavior inhibiting), their the case, then there is neither moti-
manner of implementation (for in-
vation for choice nor a basis for
stance, coercion or incentives), or
their consequences. Public policy is politics.
defined here more narrowly: instru- Second, the social system must,
to some extent, be nondetermined
mentalities adopted by governments
for the purpose of creating changes yet sufficiently integrated to be
in the states of the social system, determinable. To the extent that the
including changes in the &dquo;policy sys- system is not determined, there can
tem&dquo; itself-the total set of actors be politics. The political system to
and their relationships that deter- some extent must be an independent
variable.2 Despite the necessity for
mine governmental policy. Inducing
change includes guiding, accelerat- politics being predictive, rather than
ing, or slowing down existing pat- predicted, a substantial amount of
terns of change, an example of which policy research assumes a language
will be presented.
As public policy is purposeful, it an example of an attempt to create
1. For
involves choice. As it involves a paradigm for policy analysis, see C. W.
choice, it requires prediction, Anderson, "The Logic Public Problems:
Evaluation in Comparative Policy Research."
whether from projections, theory, or
(Paper presented to the Conference on Com-
other Theoretical or theo-
sources. parative Public Policy, Cornell University,
retically informed predictions are October 1976).
the point of convergence between 2. See, D. E. Ashford, "Policy as an Inde-

policy and social science theory. It pendent Variable or Institutions Really Do

Matter" (Paper presented to the Conference
is toward the improvement of pre- on Comparative Public Policy, Cornell Uni-
diction by the use of social science versity, October 1976).
176

of analysis in which policy and only be assessed empirically with

policy outcomes are determined by cross-county and cross-time data.
the economy or society.3 yields
A theoretical orientation
Changes in contemporary political general propositions, which when
policy analysis are required in interpreted, can predict specific
order to generate predictions that cases. As relationships among the
are testable across time and across variables differ across systems and
SySteMS.4 At the very least such change over time, it is necessary
testshave the practical advantage of to go beyond extrapolations of the
reducing the costs of partial or whole immediate past (e.g., birth rates and
Such predictions
&dquo;experiments.&dquo; school age population) to theories
will be derived from macro, theo- that explain relationships (e.g., level
retical, and structural paradigms of development of the system, as it
rather than the micro, historical pro- relates to birth rates, and as these
jections, and incremental ap- relate to educational alternatives pro-
proaches that tend to dominate em- vided by the system). Theories are
pirical analysis of public policy. general, allowing for cross-system
A macro approach addresses the and cross-time generalization; extrap-
context, including the structure of olation is system specific. Further,
the institutions that are making extrapolation is largely linear; polity
policy. For example, the urban changes &dquo;interrupt&dquo; and, conse-
centers of the United States are con- quently, are nonlinear.5
strained by the macro system rule, A structural approach requires
indeed, the constitutional principle, knowing the basic relationships in a
that there ought to be free migra- social system and how these rela-
tion of capital and labor, or of tionships influence impact. Rela-
economic enterprises and indi- tionships, such as that between in-
viduals. Such rules change over flation and unemployment, are
time; they certainly vary across sys- necessary for predicting the conse-
tems. Although the United States im- quence of policy. Determining and
poses economic barriers to indi- assessing the impact of policy
vidual migration in the form of re- within the context of different
gional pricing differentials, Poland structural rules, requires the study
also imposes political barriers, in- of structural variations across system
cluding permits for residence. Be- and time; that is, comparatively
cause these macro-system condi- and historically.
tions influence micro policy predic- A macro, theoretical, structural
tions, for example, the impact of approach to public policy analysis
housing subsidies on suburban to harden and enhance policy pre-
migration, then policy impacts can diction logically requires cross-time
and cross-system analysis. Such an
3. One stream of political science research
approach becomes compelling for
on public policy took the view that the policy highly developed systems, where
outcomes were the result of other variables. there are, by definition, more
See, for example, T. R. Dye, Understanding
Public Policy (Englewood Cliffs, N.J.:
Prentice-Hall, 1972). 5. In this
sense all policies can be viewed
4. See, H. Teune, "Comparative Policy as interrupted time series. A policy is
an
Analysis" (Paper presented to the Con- adopted at some particular point in time,
ference on Comparative Public Policy, Cor- and the "before" and "after" effects can be,
nell University, October 1976). in principle at least, evaluated.
 177

&dquo;things,&dquo; or components, that are experimental analysis of policy im-

connected as well as more &dquo;things,&dquo; pact is micro because the logic of
and, as consequence greater com- experimental design is biased
plexity of policy results as well as toward the isolatable, and control-
more rapidly changing structures.6 lable ; in this case, a specific policy,
There is an addition: levels within a specific policy target, and specific
the system. Although &dquo;levels&dquo; or the policy consequences. If experi-
organization of components is a mental principles are carried to
general phenomenon and, therefore, macro policies, those that take place
should be taken into account in the at the system level and have the
study of social systems, for policy whole system as a target-the in-
related to government, multiple dustrialization of the state-or a
levels are necessary for empirical sector of the system-the poor-
analysis of policy. With the infre- the necessary controls required for
quent exception of direct or quasi- inference become impossible to at-
military rule, all modern states have tain.9 It is one thing to determine the
a system of local government hier- impact of reading instruction on a
archically organized but with some school population by controlling for
policy autonoMy.7 class, age, community, family, etc.;
it is an order of magnitude more
Methodological considerations difficult to test the consequences of
The general assumption here is national health insurance on the cost
that social science theory can be of medical care and its distribution
used to obtain policy predictions. by studying Sweden and the United
A corollary is that it also is possible States.
to confirm predictions without the
Since the principles of experi-
cost of applying policies experi- mental design became accepted in
policy analysis in the late 1960s,
mentally to determine how they there have been few demonstratable
work.
One alternative is the successes-&dquo;’ The nonrandomization
empirical of policy treatment; that is, the fact
study of policies in force, after the
fact, one ingredient of policy evalu-
ation. At the micro level, testing pre- D. T. Campbell, "Reforms as Experiments,"
dictions according to the standards The American Psychologist (April 1969),
of evidence of experimental design pp. 409-29. See, also J. A. Caporaso and
L. L. Ross, eds., Quasi Experimental Ap-
is described with the label, quasi-
proaches (Evanston, Ill.: Northwestern
experimental design.’ But quasi- University Press, 1973).
9. This is discussed in detail in H. Teune,
6. For elaboration of the relationships
an "Public Policy: Macro Perspectives," in
between social complexity and policy prob- Process and Phenomena of Social Change,
lems, see, H. Teune, "Information, Con- ed. G. Zaltman (New York: Wiley, 1973).
trol, and the Governability of Territorial 10. See, for example, J. A. Pechman and
Political Units" (Paper delivered to the 10th P. M. Timpane eds., Incentives and Income
World Congress of the International Political Guarantees: A New Jersey Negative Income
Science Association, Edinburgh, 1976). Tax Experiment (Washington, D.C.: The
7. For an explanation of the role of local Brookings Institution, 1975). The problem
government, see J. J. Wiatr, H. Teune, and with this, as well as other experiments to
Z. Mlinar, "Alternative Local Organizational evaluate policy consequences, is that a
Policies in Development," Polish Round single experimental social context, for
Table, vol. VI (1974-75). example, one school or city, cannot be repre-
8. The parallelism between policy de- sentative of the system to which the policy
cisions and experiments was highlighted in is intended to apply.
178

that the policies that are adopted are For policy analysis these point
to a considerable extent determined predictions, however, ought to be
by factors that also influence their expressed in terms of some range
results, and the nonisolatability of of likelihood rather than as exact
the target populations; that is, the values. That is one purpose of policy
fact that target populations are em- analysis: to determine the con-
bedded in a complex social and straints and alternatives. There will
policy environment which also af- almost certainly be this much migra-
fects the consequences of any par- tion of industry from Pennsylvania
ticular policy, has limited the suc- to Texas, but if this policy is adopted
cess of quasi-experimental assess- or if this policy is strengthened,
ments of policy predictions. For in- this will be the more likely migra-
stance, did Headstart work? tion rate. Strictly determined pre-
The limited number of country diction points, of course, exclude
cases, the similarities of policies policy, unless the policy is public
emanating from industrialization, affirmation of the inevitable.
the attenuated linkages between a
national or macro policy and local THE URBAN FISCAL CRISIS IN THE
or micro effects are among the rea-
UNITED STATES: AN EXAMPLE
sons that macro policy analysis
serves primarily to sensitize aware- Rather than presenting a sugges-
ness of alternatives and possible tive inventory of theories, structures,
difficulties in implementation rather and macro-systemic characteristics
than to confirm predictions. that control and define policy op-
The lack of success of empirical tions, an example will be presented
analyses of public policy conse- in some. detail. The example of the
quences argues for a theoretical urban crisis in the United States will
approach as an alternative. A theoreti- illustrate one step in setting up a
cal approach should allow for mul- problem for macro, structural, and
tiple time point predictions or pre- theoretical policy analysis. The next
dictions that look like point predic- step, which is not pursued here,
tions.ll If it is true, for example, that would be to specify some of the
the costs of density accumulate in likely predictions that must be ob-
urban areas as a function of time tained if a theory is true. A third step
(decay as a function of city’s age) and would be to present data and analy-
if it is true that capitalistic enter- sis from several countries and for
prises are free to migrate; and several points in time to show which
if ... ,then what are the likely of the theories is more probably
rates of migration from the aging true. A fourth step would be to inter-
cities to the new cities in the United pret the theory for a specific case or
States over time? country in which a particular policy
is likely to be adopted.
11. See, for example, P. Meehl, "Theory
Testing in Psychology and Physics: A The example of the urban crisis or
Methodological Paradox," Philosophy of the urban fiscal crisis is a signifi-
Science, (June 1967). See, also H. Teune, cant one, at least for the United
"Macro-Theory and Micro Analysis: An In- States. Three macro-structural the-
terpretation of Comparative Social Research"
(Paper presented to the Seventeenth Annual ories will be presented with policy
Convention of the International Studies Asso- alternatives that may have an impact
ciation, Toronto, February, 1976). on the course of events.
 179

The urban crisis as the crisis of mand for material satisfaction and
capitalism governmental redistributive re-
In Marxist writing, one of the sponses. The latter must accelerate
in such a way that it overcomes in-
dynamics of decline of the capitalist creases in the production of material
state is the contradiction inherent in
the government’s need to maintain goods and at some point dampens
any further increases in economic
legitimacy ’through the satisfaction output as a result of investment.
of the material wants of the deprived Characteristic of many policy prob-
classes and the private sector’s need
for capital accumulation to increase lems, such as the race between the
output. This deadly dynamic states
depletion of oil resources and the
that in order for the system to sur- discovery of politically acceptable
alternative sources of energy, will
vive, the population must be pla- end up either with the incorporation
cated through welfare or transfer pay- of the private sector into the political
ments which the government pro-
one, in which case the state would
vides through taxation, and these have &dquo;transcended&dquo; the bourgeois
increasing demands for material class, or the abortion of democratic
satisfaction can be met in the long
politics and its transformation into
run only by increased investment in the tyranny of Aristotle, or the work-
the private sector.
The argument, made popular in ing class state of Lenin.
Most modern states do not have a
James O’Conner’s The Fiscal Crisis government but several govern-
of the State, 12 requires clari- ments, local governments with
fication of some macro-structural autonomous resources and inde-
conditions: (1) the presence of a
democratic government which has pendent political power. The ques-
tion for cities is: to what extent
competition for votes on the one can this dynamic of the state be
man, one vote principle and toward translated for cities and be used to
which political demands can be ef-
explain the urban fiscal crisis? One
fectively directed; (2) weakening
a
necessary condition for this theo-
of social, institutional, ideological
or
retical dynamic to become engaged
constraints on political demands; (3)
in cities in capitalistic states is that
the presence of social conditions and
the absence of political instru- they must have the possibility of
mentalities such that overt con- behaving as if they were capitalistic
flict and violence is not only pos- governments with democratic par-
sible but also a real threat to govern- ticipation or mass political move-

mental stability; and (4) elasticity in ments ; that is, they must be able to
affect capital formation in the private
the government’s capacity to tax pri-
sector, on the one hand, and to be
vate economic institutions.
The critical theoretical dynamic responsive to political demands for
increased transfer payments, on
is the race deriving from the rela- the other.
tively longer term time function be- One macro-system result that can
tween investment and increased
be derived from this is: to the extent
economic output and the shorter
term one between nolitical de-
that there is local autonomy to tax,
especially to tax private economic
12. J. O’Connor, The Fiscal Crisis of the organizations, is the extent to which
State (New York: St. Martin’s Press, 1973). the fiscal crisis of cities in capitalis-
180

tic countries will parallel those of A third prediction about the urban
the state. Conversely, to the extent fiscal crisis derives from the class
that there are limited local powers conflict origin of this deadly dy-
to tax private economic organiza- namic for cities. The extent to which
tions is the extent to which there there are class or income differences
will be less of a fiscal crisis in the within cities should predict the rate
cities relative to that of the central of increase in the urban fiscal crisis,
government. cross-nationally as well as within
In the United States, there is a countries. One factor that will com-
fiscal crisis in the cities, but a lesser pound the increase in the fiscal cri-
one at the national level. In com- sis is internal migration of lower
parison, Great Britain has relatively income population induced by better
few fiscal crises in its cities, but welfare opportunities to the cities,
a severe one nationally. The United such as from the South in the United
States has both local autonomy and States or in Italy.
political power bases focused The foregoing are illustrations of
locally and independent of the general predictions from this theory.
central government. Britain has Specific, point predictions could be
little of either. Both local autonomy made concerning actual welfare ex-
and local political power are con- penditure rates and migration of
trolling macro-structural characteris- industry out of the city or increased
tics for the operation of this theo- voting or political demonstrations
retical dynamic of capitalism within and welfare expenditures.
local governments. The policy alternatives are less
A second macro-structural vari- clear than the theoretical frame-
able that is suggestive of predic- work for prediction. One is the na-
tions about the urban fiscal crisis, tionalization of welfare, which
both within and across countries, as would depoliticize demands for wel-
well as over time, is the extent to fare in cities, thus providing cities
which there is local politics versus with time to search for other solu-
nationally focused politics. This is tions. Another would be the imposi-
more than a matter of local au- tion of restrictions on industrial mi-
tonomy ; it is the scope, intensity, gration, thus allowing those cities
and frequency of political participa- with more severe fiscal problems to
tion, in particular, that of the lower buy time. Both of these alternatives,
or relatively more deprived classes. however, would only delay worsen-
As the level of political participa- ing the fiscal crisis of the city. The
tion in cities increases over time, system is determined, at least in the
their financial condition will de- long run.
teriorate-increased taxes for trans- Historically, of course, cities did
fer payments, leading to the escape not always have fiscal crises. Since
of industry, and a deterioration of the the Depression, manifestations of
tax base, leading to even higher rates this in the United States began in the
of taxes and further reduction of ser- late 1960s.11 In the 1950s, cities
vices, leading to popular dissatisfac- experienced increasing revenues
tion. Furthermore, any substantial
increment in local autonomy, such
13. This general problem was discussed
as regionalization in Italy or devolu- earlier and formally modelled. See J. For-
tion in Britain, should accelerate the rester, Urban Dynamics (Cambridge, Mass.:
fiscal crisis of cities. M.I.T. Press, 1970).
 181

from anincreasingly productive the United States has such a crisis;

economy and, accordingly, in- Canada does not; Italy has; West
creased their revenues. In the late Germany does not. Further, within
1960s, the cities began extensively the United States some cities have a
to exercise their legal options to fiscal crisis; others do not.
borrow. The present revenue crisis Politics bears conflict, and conflict
of the cities is that they no longer brings bargains concerning col-
can tax without impairing their re- lective allocations of resources
source base. This deadly dynamic among classes, groups, regions, spe-
was engaged, of course, sooner for cial interests, and especially local
some cities than for others. governments. The urban fiscal crisis
Comparatively, because the urban in the United States is the result of
fiscal crisis is a consequence of ad- shifts in political power, following
vanced capitalism, the more ad- from a series of historical events
vanced the capitalistic nature of and the response of the political sys-
the country, the more apparent tem tothem. For example, the de-
should be this urban crisis. Also, velopment of the trolley car began
urban fiscal crisis should not be after Boston elected an Irish mayor
found in socialist countries, no and Chicago experienced the Hay-
matter how advanced their econo- market riots. The cities in the North-
mies. This conflict within the polity, east noticeably began to deteriorate
between deprived groups seeking when labor saving technologies
welfare and the private economic were introduced by government-

organizations pursuing investment, sponsored research to satisfy

does not formally exist. Southern agricultural interests.
Congress became less interested in
The urban crisis as a consequence of the cities when, after the 1960
political power census, over half of the members of
the House of Representatives came
Modern cities are one type of the from suburban areas. The South and
urban political form; they are also, Southwest became more
powerful
however, actors in national political relative to Northeast, as a result of
systems. Accordingly, one theo- increments in military procurements
retical interpretation of the urban that were
crisis is that it is an outcome of
regionally distributed by a
seniority system in Congress based
political conflicts determined by the on incumbency. Add to this, the
distribution of political power. 14 The financial burdens that followed
distribution of power must be inter- unionization of teachers and munici-
preted in the context of specific pal employees in the mid-1960s in
national political systems. There- the older cities. Consider the costs
fore, it is necessary to examine con- of imposing federal regulations on
figurations of political power of cities, such as those for clean water.
specific countries. The fiscal crisis The foregoing illustrates some of
of cities is not an inevitable charac- these &dquo;system specific,&dquo; historical
teristic of capitalistic countries: factors that converged on certain
cities to produce their fiscal crisis.
14. This is a standard type of political
The general proposition underlying
science explanation in which the distribu-
tion of resources flows from configurations this theoretical explanation is that
of political power. It takes the view that the fiscal crises of cities are the con-
politics is a form of conflict. sequence of the political successes
182

of their competitors in obtaining policies to offset the worst conse-

politically allocated resources. To quences of the urban fiscal crisis.
predict the future of cities, it is
necessary to configurate the distri- The urban crisis as a consequence
bution of political power and prob- of technological change
able changes in that distribution,
and to chart, over time, the likely The large city, as it is known in
consequences for the distribution of Europe and North America, is a his-
resources. torical phenomenon of the late nine-
Comparatively within countries, teenth and early twentieth cen-
such a political explanation would turies. The great cities developed
account for the wide variations in concomitantly with industrialization
the fiscal crisis of cities in the United and the technology supporting it. As
States. Certain cities neither have this technology has now been trans-
nor foresee such problems. Cross- formed, the large cities will decay.
nationally, this theoretical approach The urban fiscal crisis is one aspect
would explain the relative freedom of this decay.
of British, Scandinavian, and Dutch This theoretical interpretation
cities from a fiscal confrontation. It assumes that the modern city, as a
also might explain the relatively local political organization, emerged
poor state of some cities in Asian as the consequence of an economic
and South American countries, form of production defined by a tech-
where political support for the nology, and its decline is a conse-
central government comes largely quence of the replacement of its
from either rural-based masses or technological base. Just as the family
elites, and even perhaps the rela- farm, as a way of life in United
tive viability of Asian urban en- States, has almost wholly vanished
claves, such as Hong Kong and because of changes in the tech-
Singapore, where there is no com- nology of agricultural production,
petition from the countryside. which depreciated labor in favor of
Historically, such a view of the capital, so the large urban center,
urban fiscal crisis would explain the as a way of life, is disappearing. If
rise and decline of certain cities the city continues to exist in the near
within countries without regard to future, it is only because of subsidies
their level of industrialization, de- from the political system, primarily
velopment, social order, or political because of the residue, gradually re-
system. It could also help to explain ceding, of party voting among union-
why a rurally biased political sys- ized workers.
tem, such as the United States, con- Urban centers must pay the high
tinued, and to some extent con- overhead costs of density. Density
tinues, to have substantial subsidies requires more expenditures per
for rural areas, despite the decline capita for fire protection, enforce-
in the rural population. ment of regulations, and police pro-
The policy implications of this tection. These costs of density com-
macro-historical analysis of cities as pound with age; that is, the older
political actors are several. It allows the city, the greater the costs of
wide latitude for choice, or politics, density, if only because the resource
to redress the system in favor of of unused land diminishes.
urban areas and, indeed, for specific Despite’the costs of density, for
 183

at least 50 years increases in produc- tration was weakened. Add to this

tivity yielded from a coal-steam tech- the introduction of electrically

nology were greater than existing al- powered trolley cars and the auto-
Further, the coal-steam
ternatives.l5 mobile. These technologies for dis-
technology for manufacturing cre- persion contributed to the process of
ated efficiencies that not only paid distributing the population over
for the costs of necessary urban larger areas.

services, but also provided a sur- Although the rate of population

plus for investment in urban ameni- dispersion around the cities de-
ties-parks, museums, etc. 16 creased during World War II, with
Another dominant feature of the the full utilization of manufactur-
coal-steam technology is that its ef- ing capacity, it accelerated after the
ficiency rapidly depreciates with war. The economic base of the cities
physical distance, cooling the steam improved in the 1950s with the
or straining the conveyor belt. The growth in the size of economic or-
consequence of this was concentra- ganizations and governments and
tion of population around the center, the shift to service industries, which
configurating residence along class about 1952 began to employ more
and occupational lines as described than half of the working force. Large
by the classical social ecologists. 17 size organizations require man-

Small, rurally-based factories, al- agement, bureaucracy, and white

though able to discount their physi- collar workers. In the country as
cal and social pollution at zero, a whole there was a continual in-
were less efficient. The economic crease in white collar employment,
advantages of the city were not whereas jobs in manufacturing re-
marginal, but decisive. Those ad- mained relatively stable.
vantages allowed for the construc- Bureaucracies deal largely in un-
tion of urban economic infrastruc- coded information or in categories of
tures and for the social investments information that, even if standard-
necessary for the economic growth ized, are likely to change. Face-to-
of the private sector. face interaction is required to pro-
Around 1930, electricity became a vide high density communication.
viable alternative to coal-steam. Its Thus the urban centers got a respite
dominant characteristic is that it can from their decline with the construc-
be transported across considerable tion of office buildings for decision-
distances with little loss in ef- making and management which be-
ficiency. The compelling techno- gan to replace manufacturing as the
logical force for population concen- primary economic base of the
cities.l8
15. See O. P. Williams, Metropolitan Poli- But, at the same time as the
tics (New York: The Free Press, 1974). bureaucratic economy began to
16. For an example of an attempt to as-
sess the relative productivity of cities, see take shape in the cities, the com-
R. Higgs, The Transformation of the Ameri- puter, especially beginning around
can Economy, 1865-1914 (New York: Wiley, 1960 with the wide-scale introduc-
1971).
17. This is classical social ecology repre- 18. This face-to-face interaction factor in
sented by the "Chicago" school. See, for ex- the tertiary sector in the center of the cities
ample, R. E. Park and E. W. Burgess, The was pointed to in R. Vernon, The Myth and
City (Chicago: The University of Chicago Reality of our Urban Problems (Cambridge,
Press, 1925). Mass.: Center for Urban Studies, 1962).
184

tion of interchangeable languages, technology, and the structure of

again challenged the just developed human settlements, are largely con-
economic rationale of the large straints on policy choice. It is un-
urban centers. Dispersion of bureau- likely that in the long run, policies
cratic activity became not only pos- can reverse these trends. Choices
sible but also inexpensive. Payroll, can be made to control these pro-
inventory, and planning can be dis- cesses, to anticipate them, and to ex-
persed and then linked through a plore new urban forms, such as small
computer network. There is one con- linked to one another and to
towns
straint on this dispersion-the stan- few large, national urban centers.
a
dardization of information. Where The point predictions that can be
information cannot be standardized made from this theoretical interpre-
or the categories of information are tation of the urban fiscal crisis in or-
unstable, population concentration der to confirm it are several. They
for economic activity remains essen- involve shifts in the economic ac-
tial. Examples include art, publish- tivities within the large urban
ing, banking, research, and politics. centers and the concentration and
But, for unstandardized information, dispersion of economic organiza-
the need is for diversity and the tions and population.
available unstandardized informa- Historically, the growth of cities
tion is likely to be aggregated in a should be preceeded by the growth
very few national or, indeed, inter- in manufacturing. This should be
national centers.19 followed by white collar or tertiary
The foregoing takes a long-term, occupations succeeding manufactur-
macro perspective to explain the ing, followed by activities requiring
rise and decline of large cities in uncoded information. As this occurs,
industrialized countries. These the low-skilled population should
macro trends can be offset by policies become increasingly unemployable,
and other events for decades. The while, at the same time, the diversity
technological successions can take of occupations within the cities
generations. should increase.
The fiscal crisis of the cities may Comparatively, the theoretical dy-
be only one short-term indicator of namics of technology and density
these long-term processes. Short- should take place regardless of the
term changes can also be expected, nature of the political system,
such as a shift in economic activities whether capitalistic, socialistic, or
of the cities from one type of un- military rule. Here the historical
standardized information, such as and comparative perspectives can
banking, to another, such as re- be juxtaposed. If countries began to
search. Succession in economic ac- industrialize after World War II, for
tivity in large urban centers will example Eastern European coun-
be determined by the standardiza- tries, then their dominant form of
tion of information. population concentration should be
The policy implications of this to the large urban centers for manu-
theoretical approach, the nature of facturing. As the tertiary sector be-
19. Of course, the development of a few
gins to develop, manufacturing
major centers is now proceeding on an inter-
should disperse in favor of white
national scale and it is useful, perhaps, to collar occupations. These, too, then
examine the city as an international actor. will begin to disperse.
 185

A CONCLUDING STATEMENT haps, a more elaborate model for

prediction can be constructed from
The example of the fiscal crises of all of these or, perhaps, one is the
cities was presented to illustrate a dominant
dynamic for explanation.
macro, structural, and theoretical ap- In testing theories, it is surely
proach to public policy analysis. better to exclude alternatives than
Such an approach will allow for to examine one to establish its good-
predictions to specific conditions ness of fit in general. If one can be
over time and across countries. The excluded, the alternative becomes
specific predictions will provide an more credible. At the very least, it
empirical basis for confirming cer- is helpful to understand the root
tain basic relationships, which, at causes of the problem in a policy
the very least, will clarify the con- area. These causes will be theoreti-
straints on policy choice. cal, and in so far as they are, social
The three alternative theoretical theory and public policy will be
interpretations are a first step. Per- linked.