Resumé

Basically, there are two solutions to the problem of how policies determine politics. is the structural solution suggested by Theodore J. Lowi. His way of classifying policies in terms of the likelihood and applicability of coercion seems to be most valid under the general macropolitical assumption of asymmetrically organized patterns of authority. The other is the marginalist solution suggested by James Q. Wilson. His classification of policy change in terms of whether costs and benefits are being dispersed or concentrated through the change in question seems to be most fruitful when the focus is upon the logic and equilibrium a policy situation as such, regardless of its characteristics from the point of view of coercion and inducement. In this article the confrontation between and Wilson is analyzed, and a way of combining the two approaches is suggested.

One hundred years ago, people who were disposed to think about the role of public authority in society believed that it would eventually decrease. Classical Liberalism regarded such a diminution as desirable and imminent. Marxist theory of the withering away of the state gave the same notion a more dramatic formulation: through a final, gigantic harnessing of public power, society would be brought to a position where it would be able to govern itself.

As far as we can now tell, looking at developments from the vantage point of the present, it seems that both Marxism and Liberalism were wrong. Increased industrialization and urbanization have been accompanied a rise in the scope of public activities, and the model of a socialist political development seems to be far better suited to account for what is actually taking place than the older Marxist and Liberal notions.

It is unclear how this secular trend in a socialist direction relates to a far less universal trend in the direction of democracy. From the Soviet Union, China, and many countries in the Third World one hears constantly that socialist policies require a neglect of democratic procedures. Similarly, in

the West one hears constantly that democracy requires a social base of private enterprise to survive. Dominant ideologies value these two basic trends - socialism and democracy - differently. But they all regard them as impossible to achieve at the same time. Communism gives up democracy, Liberalism gives up socialism, and in the more authoritarian variants of Conservatism they are both given up. So one is easily left with the impression that the empirical issue is already settled.

Nevertheless, in Sweden, Great Britain, and some of the other democracies, in the direction of socialism have actually come so far that this old question - about the reconcilability of socialism and democracy - is being posed again, using experience as the bases for argument and having the solution of practical problems in mind. Important sectors of all these societies have come, whether intentionally or not, under public control. In Great Britain important portions of industry have been nationalized. Sweden, government has developed a system of planning and strategic oversight over the economy in general, and over land use, construction, housing in particular. Old ideological disputes over the political implications of public growth are more and more becoming matters practical concern: what kind of issues should be handled at which administrative level to ensure an optimum of effectiveness and efficiency of governmental institutions on the one hand, and of participation and representativeness in governmental decisionmaking on the other? These modern, practical discussions presuppose, in turn, a better basic understanding how politics is affected when policies are of different kinds, and vice versa.

Looking at the history of modern political science, it seems that researcherstried avoid the question of the political implications of public policy as long as possible - either because it was too controversial or because it was conceptually difficult to master. Regardless of our choice of approaches to political procedures, we can easily find a body of solid writings which describes what limited government and institutionalized opposition look like in different countries, and which gives some ideas about the conditions that cause variations in the procedural characteristics of political systems. The same could be said about the comparatively well researched problem of political participation. After several decades of intense study we know fairly well how different parts of the citizenry entered into the national political arenas; how they were enfranchised and how they gained the right to form political groups and parties. A rather well elaborated, comprehensive picture of political attitudes, organizations,and is also available, and large-scale empirical research is

being conducted on fruitfully defined problems at the level of middle range theory. The content of public policy is the third basic problem to which political science addresses itself.¹ But that literature is much less developed.It true that there is a growing interest in this problem, but most of the writings and the research seem to have a focus that is more general than a political concern with the subject matter.

On the one hand we have the kind of research design in which policies are conceived as determinants in search for its various effects. Much of the policy research starts out with a particular policy and then proceeds to explore the varied effects it might have.² Obviously, no strict analytical technique will work here, for such techniques are designed to identify and evaluate multiple determinants of a single phenomenon, not multiple effects of a single cause. On the other hand we have the kind of research design that most successful studies of policy impact have adopted, whatever initial purpose (see Figure 1). In a much discussed study, James Coleman (1966) started out v/ith the intention of determining the impact of the 1964 Civil Rights Act in I:he United States with respect to the creation of equal educational opportunities for racial, ethnic, and religious minorities. he ended up with a highly sophisticated explanation of differences far as educational results were concerned: the central focus of the study had shifted from the Civil Rights Act to the outcomes of the educational The success of the study, as an impact study, was due to the fact that Coleman had given policy the same analytical status as the other independent variables hypothetically related to the impact variable (Grumm 1975, 448 ff).

If these are the only alternatives open to the political scientist, who is

mainly interested in the political aspect of public policies, one is caught in the dilemma of having to choose between the equally unsatisfactory alternatives of focusing upon policies in search of their various effects or of focusing upon some measurable aspect of society in search of the possible impact of policy. In neither case is there any focus on the kind of questions which are central from a political point of view: what types of policy create what types of procedure and participation, and what types of procedure and participation create what types of policies.

There does exist, however, in modern research on public policy, a tentative line of argument in which the political aspect is the primary concern. Seminal among those who have tried to develop a political approach is Theodore Lowi (1964; 1970; 1971; 1972a; 1972b). In a number of publications he has maintained that the most significant fact about government is that government coerces and that 'policies determine polities'.Therefore, contends, 'the impact of policies on the political system can be predicted and planned for. The projected impact of policies on politics can be developed as a criterion of policy choice; a criterion that does not have to await the long range impact of a policy on the society.'³ Different types of coercion provide different sets of parameters, contexts within which policies take place. Accordingly, Lowi considers these types of coercion as the most fruitful criteria for classifying policies, with a

prospect of building and testing theories about their relationship with
politics (see Figure 2).

The likelihood of coercion can be more remote or less remote, according to the Lowi typology. Sanctions in the distributive case of grants, tariffs, and subsidies given away case by case are positive, and therefore mild and uncertain. The same is true in the constituent case, when the governmental administrative structures are changed. In the other two cases coercion is more immediately present; norms are imposed and resources are transferred from one part of society to another. As for the applicability of coercion, Lowi draws our attention to the fact that some policies do not come into operation until there is a question of the behavior of some specific individual. In the regulatory case there is a general rule covering all fraudulent advertising, traffic behavior and so on, but it is applicable only to the conduct of individual car drivers and advertisers. By contrast, some policies do not need to wait for a particular behavior. Re distributive policies do not touch behavior directly at all. Instead they work through the environment of conduct. For example, a minor change in the discount rate can have a major impact upon the propensity to invest. Yet no government official need know of the existence of any individual investor. No question of the legality of individual behavior is involved. Nevertheless, the interest rate to affect investment behavior is certainly more effective and efficient than urging, regulating or other techniques to adjust the environment within which individual choice is taking place.

What is so fascinating about the Lowi typology is the potential elegance and generality of the hypotheses that are implied. As soon as we put it this way, Lowi (1964,689) argues, 'it becomes clear that each kind of coercion may very well be associated with a quite distinctive political process.' These four basic types of politics, or 'arenas of power', as Lowi calls them, are all 'historically as well as functionally distinct' for the American case. Distribution was 'almost the exclusive type of national domestic policy from 1789 until virtually 1890' and it enhanced a decentralized, disaggregated 'every man for himself type of politics. The constituent issues related to the Civil War, civil rights, and the rights of the individual states brought the American two party system into existence: 'Agitation for regulation and redistribution began at the turn of the century, but regulation become an established fact before any headway at all was made in redistribution.'

But the Lowi typology is more than a scheme for the historical study of
American politics. So far, there have been mainly two interpretations of
government and politics under modern conditions. On the one hand we

have the pluralist notion of interest group bargaining and the outcome of these trials of strength being conceived as a temporary equilibrium reachedin group struggle at any given moment. On the other hand, we have the governmentalist notion of imposition of norms and reallocation of resources by a dominant public force. The applicability of these two theories, Lowi argues, follows from something that adherents of neither school of thought have made clear to themselves: the behavior of groups and parties is created by the type of policy at stake. The regulatory arena, in Lowi's terminology, corresponds to the governmental process as conceivedby pluralists and 'the school's general notions are probably found to be limited pretty much to this one arena.' The redistributive arena, in much the same way, corresponds to the governmentalist view of the process as being a confrontation of parties and organizations clustered into a few major blocs which remain stable as long as the major redistributiveissues stable and where these same blocs use the state apparatusto about their own notion of an equitable social order.⁴

For all its 'boldness and imaginativeness', however, the Lowi typology could be seen as lacking in logical rigor as well as empirical plausibility. In the most well organized attack so far launched upon the Lowi typology, James Q. Wilson (1973: 327 ff) argues that it is imperfect from a logical point of view. It is hard to distinguish between the types except in extreme cases, and there are many policies that could be classified under two or more categories. Furthermore, and of greater interest, Wilson argues that the implications of the Lowi typology are not necessarily true. Regulatory policies may often be the result of governmental dominance and redistributive may be the result of interest group politics. The latter seem to be the case when organized segments of the society make competing such as 'struggles between labor and management over the definition of unfair labor practices, between wholesalers and retailers over the maintenance of fair-trade laws, or between two firms over the awarding a television license.' But there is a broad range of other regulatory issues in which group activity is modest, or where the activity that does occur tends to be carried on by unsuccessful opponents. An example of this is the failure of concerned business groups to prevent the passage of laws beneficial to consumers. Much of that legislation, such as automobile safety, in the United States in the 1960's was passed over the objections of well-organized business interest groups (Wilson 1973, 329).

Having criticized Lowi on the grounds of logical as well as empirical
deficiencies, Wilson asserts that Lowi's basic idea-that the substance of a
policy might tell us something about the kind of politics associated with it -

seems correct but that the conceptual scheme based on that insight requiresmodification at least two ways. The first modification Wilson suggests is that we distinguish between 'the adaptation of a new policy and the amendment of an existing one.' The most important new policies of government are adopted only after there has been a change in opinion or a new perception of old arrangements sufficient to place on the public agenda what once had been a private relationship, and to clothe a particularprogram legitimacy. Organized groups can rarely accomplish unaided such changes in opinion or such redefinitions of what constitutes legitimate public action. These changes are 'the result of dramatic or critical events, extraordinary leadership, the rise of new political elites, and the accumulated impact of ideas via the mass media of communication.'Revisions existing policies, on the other hand, follow a different pattern that in turn depends on the extent to which the initial policy decision settles the ideological and normative issues and on the distributionof and losses entailed by the policy. Most of the new and enlarged powers acquired by government are soon taken for granted, and the debate over the propriety, not of their success, is stilled. What is true about the politics of a nev/ policy is not necessarily valid for the politics of amending that same policy at a later stage of its development (Wilson 1973, 350).

The second and more far-reaching of the modifications Wilson proposes is that 'the incidence of costs and benefits of a policy should not be obscured by the use of categories . . . that are hard to define and to purge of misleading implications.' He therefore suggests the replacement of the Lowi typology with a new classification of policy on the basis of'whether the cost and benefits are widely distributed or narrowly concentrated from the point of view of those who bear the costs or enjoy the benefits.' For example, general tax burdens, rising crime rates, the extensive sale of obscene literature are widely distributed costs. By contrast, a fee or impost paid by a particular industry or locality, or a highway construction program that destroys a particular community represents narrowly concentrated Similarly, a benefit may be widely distributed, as with social security and unemployment compensation, or narrowly concentrated, as a subsidy paid to a particular product, prestige conferred on a person or group, or a license to operate a television station. Whether costs and benefits are tangible or intangible is of no importance, and a widely distributed benefit may or may not be what economists call a collective good (Wilson 1973, 331).

This leads to an alternative typology (Figure 3) of a very simple and

elegant structure which Wilson (1974) has used in a study of the politics of regulating the pursuit of private business in the United States. There have been, in general, two main theories of the causes of regulation. In one, regulation results when legislators, mobilized by a broad social movement or energized by a dramatic crisis, enact laws designed to prevent a firm or industry from carrying on certain practices. Wilson describes this as the 'public-interest' theory of regulation. The other, currently more in fashion,is regulation results when an industry successfully uses its politicalinfluence obtain legal protection for itself or to impose legal burdens on its rivals. This is the 'self-interest' theory of regulation. The origins of many regulaory laws are hotly disputed by adherents of each theory. By secondary analysis of case studies Wilson (1974, 137) seeks 'to explain, insofar as available facts permit, the circumstances under which regulationbecame possible, the patterns of constraints on legislative behavior that are likely to emerge from a given political context, and the forces that will influence how a regulatory agency does its job.'

In the first type of situation, where both costs and benefits are concentrated where two clearly defined, undifferentiated sectors of the society contend over the allocation of costs and benefits), regulatory constraints point in the direction of creating a contract, agreement or charter which contains a definition of the competing rights and obligations of each party. No one organized sector will be able to dominate permanently administrative arrangements created to implement such agreements. will be continuing efforts to renegotiate or amend the contract, the visibility of the issue will be relatively high because there is conflict. Each part will attempt to enlist allies, and there will be frequent appeals to the courts and their functional equivalents in other political systems (Wilson 1974, 142).

In the second type of situation, where benefits are concentrated but

costs widely diffused (i.e., a small, undifferentiated beneficiary can receivesubstantial by imposing small costs on the population at large), regulatory constraints are in the direction of creating what is best called a monopoly. The beneficiaries seek to eliminate or reduce price competition within the affected industry, Entry will be made more restricted, or at least more expensive, and the regulated industries will strongly influence the regulatory agency that administers the policy for the sector of industry concerned. The industry and its agency will seek to maintain a position of low visibility to avoid stimulating the formation of an organization representativeof who bear the costs of the regulation, and, should the regulation become controversial, it will be defended by attempting to show that eliminating price competition is an appropriate means for ensuringsafety, fraud, and promoting amenity (Wilson 1974, 141).

In the third type of situation, where costs are concentrated and benefits widely distributed, regulatory constraints point in the direction of creating what I would call issues. In democratic politics parliamentarians have much to win and little to lose by focusing public attention on such questions. are, therefore, likely to ensure vital publicity and develop considerable momentum in and around the legislature. Proposals related to issues usually focus on an 'evil', personified if possible in a corporation, industry or victim, and there is little incentive in the developmental process to accommodate conflicting interests and thus little incentive to find a politically acceptable formula which all affected parties can live with: 'To compromise the proposal would be to sacrifice the capacity of the bill to mobilize support by its moralistic appeal . . . [T]he proposed solution to the problematic business practice will be shaped as much by the political process by which the proposal is generated as by an analysis of the problem itself (Wilson 1974, 146).

The fourth type of situation is of little interest from the point of view of conflicting interests. It is therefore set aside in Wilson's empirical study. But for the sake of an exhaustive presentation in the present context, it may be predicted that regulatory constraints under those circumstances point in the direction of creating a social agreement best characterized as a situation of insurance. If both costs and benefits are widely scattered, there is no reason why a political issue should evolve, except for, possibly, the influence of what political ideologies have to say about the legitimate scope of public action.

From a purely logical point of view there are few political typologies as
clear as that with which Wilson presents us. The problem in his case comes
with the way it should be applied.

A question which keeps puzzling me - and where Lowi and Wilson should have something to say - Is the relative absence of controversy in Swedish politics over policies which are highly redistributive. The usual way of dealing with this problem is to stress the homogeneity and pragmatism Swedish political culture. But that begs the question, as I see it. We need a more substantive theory that tells us why current controversy is marginal in kind, despite the fact that enormous resources are being redistributed every year through taxation and the establishment of equal access to medicine, education, meaningful work, decent housing, pensions, so on. What are the underlying dimensions and relationships which are operating here?

In the light of this empirical question, there are, in my view, two weak
points in Wilson's application of his basic scheme and, by inference, in his
argument for the abandonment of the Lowi typology.

First, Wilson does not consider whether costs and benefits are to be measured relative to a state of equilibrium or whether they should be gauged against a situation of non-intervention. In Swedish taxation policy, for example, the issue is often of this kind: Should the marginal tax rate for people in the bracket of 50-60 000 Sw.Kr. be 60 per cent rather than 65 per cent? From the point of view of scratch, where taxes are proportional or where there is no income tax at all, tax certainly takes from the few and gives to the many. But the issue at stake is conceived by the actors themselves as granting a benefit to a few people at the expense of the widely distributed cost of the majority.

Wilson deals with the problem when he asks that we should distinguish between a 'new' policy and the 'amendment' of an existing one. But in a more developed socialist system, where situations of nonintervention are rare exceptions, and where a delicately balanced, complicated system of distributive, constituent, regulatory, and redistributive measures is the prevailing pattern, the question of what is a new policy and what is the amendment of an existing one becomes primarily a matter of time perspective. The distinction hence seems to be of little help; the problem is still there.

The other weak point in Wilson's application of his cost-benefit scheme is how he handles the problem of the basic unit of analysis. He suggests that 'policy change', rather than policies themselves, should be the object of examination, and he sticks to that notion when analysing what political pattern accompanies the introduction of business regulation along differentlines. what about policies which do not change? Can we exclude from our analysis that steady bulk of public policies that is a growing

continuation of past governmental activities and still claim that we are
studying the political implications of public policies?

The answer to these two objections should be sought, it appears to me, along the same line in which it was once found in the history of economic analysis, i.e., (1) by keeping static and dynamic considerations apart, and (2) by looking at marginal changes from a static equilibrium as something analytically distinct from the substantive, structural characteristics of the way citizens are being induced by the policy in question. This has nothing to do with one's interest in the historical development of that particular policy. For, as Schumpeter (1954,964) puts it, 'statics' and 'dynamics' are independent of'stationary' and 'evolutionary' states: 'We may describe a stationary process by a dynamic model; this will be the case whenever we make the conditions for stationarity of a process in preceding periods. We may also describe an evolutionary process by a succession of static models: this will be the case whenever we deal with disturbances of a given state by trying to indicate the static relations obtaining before a given disturbance impinged upon the system and after it has had time to work itself out.'

Interpreting Wilson's suggestion in the light of Schumpeter's set of distinctions, variation as to policy change in relation to a stated equilibrium one basic way of approaching public policy. This mode of analysis could be applied regardless of magnitude, speed, duration, and political significance of the change in question. Whether a change is major or merely an amendment seems to be primarily a matter of historical and comparative perspective.

Interpreting Lowi's suggestions in the light of the same set of distinctions, as to the imposition of policies is another equally important politically relevant way of approaching public policy. Policies may be surrounded by conditions that are very idyllic from a political point of view as well as by situations which are filled with strife. Change of policy may be an easy way of avoiding as well as facing political problems. The imposition of a policy may evoke a considerable amount of conflict because of the very fact that goals, strategies, and institutions are decisively When one is Interested in variations of policy imposition, one is not interested in public policy from the point of view of change in policy but from the point of view of change in society. One focuses upon how citizens are being induced through distributive, constituent, regulative, redistributive measures and how that, in turn, reacts on the political system and vice versa.

It is from these two different points of view that Lowi and Wilson look

upon public policy without making that perfectly clear to themselves. Their respective thinking in different terms is not primarily a matter of more or less fuzzy concepts. Nor is it a matter of formulating the same basic problem in two different ways. I would argue rather that together they have established what should constitute the two basic concerns of policy research within the framework of political science. Wilson's classificationof change in terms of costs and benefits and their relative concentration and distribution seems to be most fruitful from amarginalist point of view; i.e., you focus upon the logic and equilibrium of a policy situation as such regardless of its, characteristics from the point of view of coercion and inducement. By contrast, Lowi's classification of public policies from the point of view of imposition seems to be most valid under the structural and macropolitical assumption of asymmetrically organized 'patterns of authority' (Eckstein 1973). We then focus upon variations in the way citizens are being induced, because we want to know what difference that makes.

As soon as we have made the distinction between these two approaches, it becomes clear that Wilson is too rash in his negative judgment about the Lowi typology. The distributive, constituent, regulative, and redistributive are not as clear as they should be from a logical point of view. But I cannot see why definitions could not be worked out. The important point that I want to make is that the structural categories should be kept analytically distinct from the marginalist categories with which Wilson provides us. For, by keeping these two basic sets of distinctions apart, we could check whether policy structure is more important than the marginal change at stake in any given policy situation. The problem thus is not whether we should look at policy changes from the point of view of equilibrium or from that of the substantive structure of inducement but rather ask: (1) what difference does it make from the point of view of procedures and participation whether imposition is distributive, constituent, or redistributive? and (2) what difference does it make from the point of view of procedures and participation how marginal costs and benefits are distributed?

My own tentative answers to these two questions are that (1) procedures and participation are structured as Lowi has it, in accordance with the way public policies are being imposed, but that (2) the intensity of the actual controversies correlates to the political diffuseness of the marginal issues at stake. If these hypotheses are true, they make a case for the possibility of reconciling socialism and democracy. Issues of taxation, the gradual erosion of managerial prerogatives, and the continuous growth in basic

welfare policies are not being fought with the intensity predicted by the policy structure already established. What matters from the point of view of intensity of political conflict is the characteristics of the agenda, and, practically speaking, the paice with which costs and benefits are redistributedand skill with which political coalitions are upheld and gradually restructured. With these two basic questions formulated and suggestions made as to what the correct answers might be, the problem then is how to define the basic concepts arid how to proceed in testing the correctness of my two assumptions.

The policies I have in mind are Swedish policies concerning land use, construction, and housing. These are the sectors where public growth has been most remarkable over the last fifty years, and which are broad and complicated enough to provide a testing ground for every conceivable problem implied in the idea of keeping the structural and the marginal questions apart. If a combined structural and marginal approach does not work in the context of physical planning, public housing, and redistribution housing costs, it is hard to imagine where it would otherwise work. For it can certainly be argued that policies towards housing, construction, and land use are at the very core of public concern in Sweden, both structurally and as far as continuous marginal changes are concerned.

Following Lowi, it seems appropriate to begin by identifying what a policy is not. Policies toward land use, construction, and housing are not merely statements of problems. The fact that in 1975 236,857 out of 3,324,956 household dwellings were overcrowded is not a policy.⁵ In the same way, sentiments expressing grievances about the lack of regional balance are not policies. At best, documentation and sentiments are assertions that a policy dealing with the problem is considered as desirable.Neither a mere appropriation of money to back a sentiment a policy. Statements of facts, coupled with expressions of sentiment, and attached to an authorization for expenditure may be typical elements of many public policies. But they are certainly neither efficient nor effective unless they are accompanied by some sort of authoritative analysis and a setting of priorities. It could be said that there is no policy without 'a general statement by some governmental authority defining an intention to influence the behavior of citizens by use of positive and negative sanctions'(Lowi 286). But not even this is enough as a basic definition. A policy can influence behavior by use of monetary subsidies to tenants renting a modern apartment at a higher cost rather than renting an old one at the same or lower cost. Or, a policy can seek to influence conduct by threat of punishment, as in the case of regulating pollution and littering. A

single decision, a specific issue, or a description of the activities of a public official all constitute 'the data by which we instruct ourselves about a policy.' These are not policies, however. From a structural point of view public policies must possess at least the following five characteristics: (1) an overt intent: an official expression of intentions concerning a desirable conduct or state of affairs; (2) a provision for inducements, positive, negative, or both; (3) programmed resource flows as reflected in budgets, plans, and administrative arrangements; (4) actual resource flows: patternsof must actually occur and not only remain programmed commitments; and (5) a certain minimum of effectiveness and performance:public must actually have some impact upon society - there must be some 'outcome' as distinguished from mere outputs ⁶.

From this it follows, it seems to me, that it is not the policy 'itself - whatever that means - but variations as to the relation between government citizens that should be classified. Paraphrasing Dahl's definition of 'influence', I suggest we conceive 'policy' as a relation between government citizens in which government induces citizens to act in some way they would not otherwise act.

The basic relation between government and citizens is 'constituent'. Through their very existence governmental institutions do not induce citizens to do anything in particular. Nevertheless, the constituent relation makes a considerable difference; it is a necessary, but not a sufficient prerequisite for every other relation between government and citizens. For purposes of physical planning and manipulation of housing costs, questions of territorial and administrative jurisdiction are extremely important. as well as citizens, communes, public housing authorities, housing co-operatives have to be perfectly clear about the territorial and administrative boundaries and what kind of questions have to be handled at each administrative level.

In addition to a basically constituent relationship there may be distributive,regulatory, redistributive relations between government and citizens. Relations are 'distributive' when government induces each single case positively, i.e., when it refrains from punishing misconduct and concentrates upon rewarding what it defines, case by case, as good conduct.Marginal for building research - in contrast to the basically 'constituent' organization of universities and research institutes - are of this kind. And so are appropriations for rebuilding and insulating older houses. Government considers research and insulation as good things. But it considers it as inefficient and ineffective to punish misconduct.Instead

duct.Insteadit concentrates upon stimulating what interested citizens
tend to do in their own self-interest.

'Regulative' is when government prescribes negative sanctions for not adjusting to an authorized norm by taking each single citizen's deviant behavior into consideration. Traffic regulation may provide the paradigmatic By actively punishing traffic offenders, the government demonstrates the limits of trusting citizens' goodwill and fear for their lives. It cannot induce good driving behavior merely by rewarding good conduct or relying upon redistributing wealth, income, and knowledge in order to induce citizens to become better car-drivers. As far as land use and construction are concerned, regulation is the basic method of inducement.

'Redistributive' is the relation when government manipulates the environment choice by withdrawing or supplying resources which either strengthen or weaken individual citizens' positions in the market place without taking any single person's deviant behavior into consideration. Taxation and interest rates provide key examples on the negative side. Taxes and interest rates do indeed induce citizens into a more frugal behavior in the market place than their conduct would otherwise have been. On the positive side pensions, allowances for children, tax deductions family houses, and provision of public housing are factors of great importance for citizens' actual behavior, even if government refrains from punishing misconduct or rewarding good conduct by taking individual choices into account. One may live wherever one wants and pay whatever rent one can afford. But taxes, interest rates, and governmental provision of public housing provide at structure, within which one is very much induced when one makes up one's mind.

On the marginal side of our analysis, 'costs' and 'benefits' seem to be clear enough concepts as far as political science goes. The dichotomy between 'more distributed' and 'more concentrated' is not that clear. It could be interpreted in two different directions. Obviously, it could mean that costs and benefits are becoming spread over a larger part of the population or concentrated in a smaller part. But it could also mean, in a way that is relevant from a political point of view, that they are concentratedin part of the population that is more clearly distinguishable, however much larger in size, or distributed to a part of the population that is more diffuse, however smaller in size. For more finely tuned analyses of individual cases, this might very well turn out to be an important and fruitful distinction to make. So I suggest that we build this possibility into our definition by stating: (1) a cost or benefit becomes more narrowly

'concentrated' when those who bear the costs or benefits become fewer and more clearly distinguishable, and (2) a cost or benefit becomes more widely 'distributed' when those who bear the costs or enjoy the benefits become a larger and more diffuse group of people. Whether the combinationsof and 'more distinguishable' on the one hand, and 'smaller' and 'more diffuse' on the other, actually occur will then remain an open question; it is not settled by definition. Through the definitions we have chosen, we are kept aware of this potential distinction. But we should not react until we have seen actual cases where it might apply.

In any public policy, inducements are to a varying degree distributive, constituent, regulatory, and redistributive. That gives us one basis for classification of policy situations. On the other hand, every policy situation be thought of in terms of change - real or potential - and that gives us the marginalist classifications in terms of costs and benefits. The problem of political implications of public policy could then be restated as a matter of description and analysis of a variate as represented in Fig. 4.

A conspicuous amount of change in the constituent relationship betweengovernment citizens in Sweden is easily demonstrated by comparing administrative divisions over time. The most important of these changes are those in the number of communes from 2,500 in 1951 to 277 in 1979. Lowi would predict that the making of this policy would be centered in the parties and in the top officialdom, thereby excluding the possibility of any significant amount of interest group activity. This looks like a reasonable prediction. Both costs and benefits of the reorganization were probably diffused enough to turn most of the policy situation into examples of case 1. But one cannot neglect the fact that there were some

fairly distinct groups of people upon whom costs and benefits were concentrated.There some 'losers', the amateur politicians in the small communes fulfilling administrative tasks, and there were some 'winners', school teachers, doctors, communal engineers, social workers, librarians and so on. The centralization and professionalization of communal administrationwas favor of these latter groups. So there were probably some aspects of cases 2, 3 and 4 as well.

In regional development policy the relation between government on the one hand and firms, entrepreneurs, and communes on the other is distributive. Parliament approved the 2,500 million SwKr program in 1973, it was explicitly stated that only stimulation measures were called for. Regulative and redistributive means were not to be applied, and the market forces were to be interfered with as little as possible. In practice, this meant that individual firms and their development plans had to be dealt with case by case. For support is not to be granted to projects which may be detrimental to existing firms in the areas, and only areas with a good developmental potential in the long run can count on obtaining support. From a political point of view it must make a considerable difference how evenly these benefits are distributed. In case 5 politics might very well be of the every-man-for-himself type that Lowi predicts. But in cases 6, 7, and 8, where either costs or benefits are concentrated, it seems worth looking for strong factions within existing organizations and parties pressing for policy change in a direction more favorable for the type of region they represent.

Wilson's findings about charters, issues, and monopolies provide good starting points for any further study cases 10, 11, and 12. The same could probably be said about case 9, the insurance case. Lowi predicts that regulative politics should be dominated by bargaining interest groups, and that this kind of conflict should be somewhat more decentralized and disaggregated than in distributive politics, but less decentralized and aggregated than in redistributive politics. It is certainly worth testing whether this idea holds when one controls for the distribution of costs and benefits in situations of changing public regulation.

A set of good cases for studying a regulative policy under different conditions of costs and benefits is provided by Swedish policies towards physical planning. The basic right of a landowner to build on his own land was abolished by legislation in 1947 as far as densely populated areas were concerned. What is now at stake is how and to what degree this same idea should also apply to more sparsely populated areas. Decisions concerning the management of land and water resources were taken in 1972 and

include guidelines for planning the management of natural resources in particular areas and for indicating how particular activities should be treated in physical planning. The implementation of these guidelines will be divided into two stages: a programming stage and a planning stage. After a consultation procedure, the government will make its decision concerning these proposed programmes. Upon its completion- by the end of the 1970's -a nation-wide land use plan may be said to be in existence. The process of regulating land use on a nation-wide scale should, from a distinctly political point of view, contain many clear cut cases where costs and benefits are widely distributed or narrowly concentrated in different combinations.

If we then proceed to the redistributive type of relations between government and citizens, we find such relations at the very core of Swedish housing policy. Fifty to sixty per cent of new dwellings constructed year are built by public bodies and housing cooperatives, and eighty to ninety per cent are financed with the help of subsidized government loans. Another way of presenting this is to say that in the government budget for 1975/76, 1,345 million SwKr were voted for housing and 2,150 million SwKr for subsidizing interest rates. On the other hand - and this is politically very important in the Swedish context - some 2,000 million SwKr are given in 'subsidies' to people owning one-family houses. That is the practical effect of tax deductions for interest rates in a system of steep income tax progression. The manipulation subsidies, tax deductions, and housing allowances within the framework of a redistributive structure of inducement requires a more subtle analysis and a more intuitive sense for marginal political effects on the part of the well-organized actors than any other issue that one can think ofin Swedish domestic politics.

However demanding housing policy might be from a political point of view, when one studies the debate over the issues at stake in this kind of politics in Sweden, one is struck by the relative absence of political concern by comparison with the abundance of sociological and economic considerations. Urban segregation is the problem as the sociologists conceive and the actual distribution of housing costs and gains from capital formation in the housing sector is what economists are arguing about. The political question of how to build coalitions through democratic politics in order to achieve some actual change in urban segregation and actually re-distribute costs and capital gains between different parts of the population dealt with only intuitively by the politicians themselves, and never by the political scientists.

This, in turn, seems to be part of a larger problem. Public discussion over redistributive issues in modern democracies seldom comes closer to the more subtle trade-offs of political feasibility than some vague allusions to the 'complexity' of the situation and the 'irrationality', 'incompetence', and 'talkativeness' of elected representatives on the one hand, and some sinister ideas of their extraordinary powers in conspiring against 'the real interests' of the electorate on the other. For those of us who are sensitive to a certain form of 'post-modern' political vulgarity and a slightly antidemocratic these allegations must be interpreted as a sign. For it might very well be that increased modernization - paradoxical as it may seem - is undermining the preconditions for democratic procedures (Huntington A development in that direction can hopefully be counteracted. to put it in straight normative terms, it simply must be counteracted. For there is no way back to plain Liberalism, and hopes for a Communist stateless society do not seem to be particularly well founded.⁷ Political modernization and the continuation of public growth are processes seem to be practically irreversible and untranscendable.

This last remark brings me back to my overall perspective. The political implications of socialist policies in a system of democratic procedures can certainly be described and explained as a growing complexity both in terms of policy structure and in terms of the marginal redistribution costs and benefits in situations of change. But there is no reason why political scientists should let boredom with political complexity foster an antidemocratic when we know what the only remedy is - a systematic of the political as opposed to the social and economic implications of public policy.

---

NOTES

1 For the idea of a three-dimensional classification of political regimes in terms of participation, and policies, see Blondel 1969, pp. 35 ff.

2 For a good overview see Grumm 1975.

3 Lowi 1972b pp. 287 ff. (cit. p. 289) and Lowi 1972app. 298 ff. 'One might quibble', as Lowi puts it, 'over fine points and special definitions, but, as a general rule, it is wise to follow the proposition that it is impossible to have a government policy without having coercion.' (Lowi 1972b, p. 287). That is not to say, however, that 'inducement' and 'imposition' might be better terms, referring to the 'stick', as well as to the 'carrot'.

4 Lowi 1964, pp. 695 ff: on pp. 692 ff. Lowi relates the findings in Schattschneider 1935 on American tariff policies to the distributive arena and considers that as a third major type of politics. 'Governmentalist' is the word I prefer to Lowi's 'elitist' (Lowi 1964, p. 680), which tends to suggest associations that are irrelevant in this context. The primary concern is to contrast 'pluralism', not to discuss the issue of 'elites' vs. 'mass'.

---

4 Lowi 1964, pp. 695 ff: on pp. 692 ff. Lowi relates the findings in Schattschneider 1935 on American tariff policies to the distributive arena and considers that as a third major type of politics. 'Governmentalist' is the word I prefer to Lowi's 'elitist' (Lowi 1964, p. 680), which tends to suggest associations that are irrelevant in this context. The primary concern is to contrast 'pluralism', not to discuss the issue of 'elites' vs. 'mass'.

5 According to the 1967 norm, under which the number of occupants in each apartment is not 'allowed' to exceed two per room, exclusive of kitchen and one living room. According to the 1947 norm, under which people were supposed to use their living-room as bedroom as well, 16,562 households were overcrowded in 1975. Statistisk Årsbok 1979, p. 241.

6 The second point is adapted from Lowi 1972b, p. 286, and the other four from Anderson 1973, p. 24.

7 Samuel Beer (1973, 79) argues that the secular trend 'does not necessarily imply that a greater proportion of needs and resources will be politicized - that is, there will not be an inherent (my italics) trend in modernization toward socialism.' That might very well be. But there is very little empirical evidence to support this view. For a good overview of the studies on the actual growth of the public sector, see Tarschys 1975.

REFERENCES

Anderson, C. W. 1973. 'System and Strategy in Comparative Policy Analysis: A Plea for a
Contextual and Experential Knowledge', Paper, APSA annual meeting.

Beer, S. H. 1973. 'Modern Political Development' in S. H. Beer and A. B. Ulam, eds.,
Patterns of Government. 3rd ed. New York: Random House, pp. 3-115.

Blondel, J. 1969. An Introduction to Comparative Government. London: Weidenfeld and
Nicolson.

Coleman, J. et al. 1966. Equality of Educational Opportunity. Washington, D. C: Office of
Education.

Eckstein, H. 1973. 'Patterns of Authority: A Structural Basis for Political Inquiry', American
Political Science Review 67, 1142-61.

Grumm, J. G. 1975. 'The Analysis of Policy Impact' in F. I. Greenstein and N. W. Polsby,
eds., Handbook of Political Science. Reading, Mass.: Addison and Wesley. Vol. 6, pp.
439-73.

Huntington, S. H. 1974. 'Post-Industrial Politics - How Benign will it be?' Comparative
Politics 6, 163-191.

Lowi, T. J. 1964. 'American Business, Public Policy, Case Studies, and Political Theory',
World Politics 16, 677-715.

Lowi, T. J. 1970. 'Decision Making vs. Policy Making: Towards an Antidote for Technocracy',
Administration Review SO, 314-25.

Lowi, T. J. 1971. 'The Development of Arenas of Power' in O. Walter, ed., Political
Scientists at Work. Belmont, Calif.: Duxbury Press, pp. 85-105.

Lowi, T. J. 1972a. 'Four Systems of Policy, Politics and Choice', Public Administration
Review 32, 298-310.

Lowi, T. J. 1972b. 'Population Policies and the American Political System' in A. E. Keir
Nash, ed., Governance and Population: The Governmental Implications of Population
Change. Washington, D. C: Government Printing Office, pp. 283-300.

Schattschneider, E. E. 1935. Politics, Pressures and the Tariff. New York: Prentice-Hall.

Tarschys, D. 1975. 'The Growth of Public Expenditures: Nine Modes of Explanation',
Scandinavian Political Studies 10, 9-31.>

Wilson, J. Q. 1973. Political Organizations. New York: Basic Books.

Wilson, J. Q. 1974. 'The Politics of Regulation' in J. W. McKie, ed., Social Responsibility
and the Business Predicament. Washington, D. C: The Brookings Institution, pp. 136-68.