Resumé

By using game-theory, it is indicated that encompassing organizations may participate in dilemma games with free-rider problems. Next, examinations of institutional details of Norwegian corporatism out that creative or productive strategies may be quite likely even among small interest organizations. In the last section it is argued that the assumption of economic growth being a collective is more acceptable the larger the public sector. Again, institutional details need to be in focus. The model closing the paper dwells on whether comprehensive organizations really have stronger incentives to choose creativity rather than to fight for redistributions. A general affirmative answer cannot be provided.

Introduction

The Norwegian 'Power Research Project' presented its final report in the autumn of 1981, concluding that special interest organizations had increased their influence society. The triple alliance between special interest groups, government and Parliament committees provides the main vehicle for wielding this influence.

At the same time, Mancur Olson published his Rise and Decline of Nations (1982). He suggests that special interest group activity will lead to an undersupply collective goods such as economic growth. The resources are used to increase the share of the pie rather than increasing the pie itself. Thus far Olson's reasoning is in accordance with that of the Norwegian power researchers.

Olson goes a step further, however, and posits that encompassing interest organizations will have an incentive to support growth-geared policies. The relatively Norwegian economic growth is to be explained by the encompassing nature of such coalitions. According to Olson, the collective action problem is then less likely to arise.

In the next section, we present some game-theoretical interpretations of Olson's argument. The analysis yields conclusions that are only slightly different from those found in The Rise and Decline of Nations. The third section dwells on economic development and aspects of the interest group system in Norway. In the last section it is discussed whether economic growth virtually is a collective

---

* We would like to thank Jon Hovi and AJfred Ehrenclou for helpful suggestions.

good - if so, comprehensive organizations may have greater incentives both in
supplying the collective good and in increasing their share of the 'pie.

The Distributive Strategy and Institutional Details

Olson's concern is the economic, social and political effects of distributional coalitions or special-interest organizations. One of the implications drawn in The Rise and Decline of Nations is that distributional coalitions reduce the rate of economic growth in a society. A crucial question then becomes: What exactly are distributional coalitions? It is worth emphasizing that Mancur Olson's independent is a rather crude dichotomy. 'Distributional coalitions' is shorthand interest groups and coalitions that act distributively. The distributive strategy - to use a concept from game theory - is a rather narrow one; distributional strive to obtain a larger share of the society's production for their own members. On the other hand, what is termed 'creative' or 'productive' strategies are also available to interest groups. That is, organizations may try to act in the interest of the society in which they operate, for example by helping to create income and raise productivity rather than to lobby for redistribution. Now, Olson claims that small interest groups tend to choose the distributive strategy. Small in this context means small in relation to the economy or society as a whole. Only encompassing or inclusive organizations have some incentives to act for the best of society.

Olson has developed a parsimonious and precise theory. When it comes to realism accuracy of implications, the scores seem to be lower. The theory rests to a large extent on the reasoning concerning incentives presented above. Olson's discussion on this point is, however, almost devoid of institutional details; nuances with regard to arrangements of interest representation and decision mechanisms are left out. Institutional details may, however, be quite important if we want to grasp the incentives actors face, and hence, their strategic choices.

Before we turn to characteristics of the Norwegian polity, Olson's theory may be cast in game theoretical terms. With a large number of small organizations in a society, a 'traditional' problem of collective action exists. The interest group interaction can be modelled as a Prisoner's Dilemma. Each interest group has a dominant strategy; the distributive strategy (D) is better than the creative strategy (C) regardless of how other groups act. The case of encompassing organizations is more complicated, and more truly 'strategic'. There will be several encompassing in a society - if any at all. Moreover, because the behaviour of an encompassing organization may affect society's productivity considerably, it is not likely that such organizations will (generally) have D as dominant strategy.

The members of the highly encompassing organizations own so much of the society that they have
an incentive to be actively concerned about how productive it is; they are in the same position as

a partner in a firm that has only a few partners (Olson 1982, 48).

Leaving institutional details aside, it is not at all clear what kind of preference structure the encompassing organizations will possess. One possibility is that strategy C dominates strategy D, and this is the most reasonable interpretation of Olson's own position. If the society in question is characterized by very few and very large organizations, we can probably expect that strategy C dominates strategy D. But, as the: number of organizations in the society increases, freerider certainly emerge. Consequently, there need not be very many organizations in the society before other preference structures than the one presented become likely. At this point it should be emphasized that, in advanced democracies, there are normally several large peak associations and a large number of small unaffiliated interest groups. Thus, a preference structure worth considering is the game of Chicken.

In the Chicken game (see Fig. 1), free-rider problems are present. Contrary to Prisoner's Dilemma and the 'traditional' collective action problem, however, strategy C is the favourable choice if none or only few other encompassing organizations peak associations choose C. Because of the great impact each actor may have on society's productivity, it is reasonable to expect that to be the sole cooperator is preferable to defecting along with everyone else. If many - or at least some - other organizations behave in a social way, the encompassing organization question may find it most favourable to choose distributively - or to free-ride. In situations where a large part of the society acts creatively, the benefits encompassing organization can gain from defection may be higher than benefits that could have been reached by cooperation or creative action.

Under reasonable assumptions, Chicken games have several equilibrium points; each actor prefers an equilibrium where he chooses D to equilibria where he is member of the creative or productive 'coalition'. It is more important, however,that solution of the game cannot be predicted. No actor has any reason, at the outset, to choose one strategy rather than the other one. Each strategy

choice is highly dependent on information about others' preferences and likely choices. An encompassing organization behaves productively if its leadership or members believe that others to a great extent behave distributively. On the belief or guess that other actors choose strategy C, the same encompassing organizationwill distributively. As it turns out, the game-theoretical interpretationis firm ground for drawing precise and testable implications.

We do not think that the game theoretical model is inconsistent with Olson's theory, it just complicates the case of large or inclusive organizations. Note that Olson is careful in formulating his implication 5 on encompassing organizations. refers 'to the incentives that face encompassing organizations rather than to their choices in particular cases' (Olson 1982, 53, emphasis added).

Norwegian Organizations and Institutional Mechanisms

For the case of Norway, the development in GDP and percent annual growth of GDP are depicted in Fig. 2 and 3 respectively. The time period 1865-1985 is covered. Since the Second World War, the growth of GDP has been rather high, however, with a minor setback immediately after 1960. As Fig. 3 shows, the percent growth of GDP from year to year has with few exceptions been positive over the whole time periode.

Fig. 4 illustrates the development of total union-membership in the Norwegian
of Trades Unions (LO: Landsorganisasjonen i Norge).

Since the establishment of LO in 1899, its membership has grown throughout the period, apart from two setbacks during the world wars. Also, the diagram shows that membership growth culminated during the sixties. Recent development, indicates an increased growth of membership.

Fig. 5 illustrates the development of membership in LO relative to total employment. diagram yields a picture quite similar to that of Fig. 4. We may note, however, that since the late sixties, union-membership in LO has culminated to total employment. This is partly due to an exceptional growth in the labour force. Also, trade unions outside the LO have doubled their membership from well over 100,000 members in the mid-sixties to almost 250,000 members in 1980. Thus, the dominance of LO seems to have diminished somewhat during the two last decades.

The observation that Norway's growth rate during the seventies was quite respectable in comparative perspective, despite the multitude of interest groups in the Norwegian society, is explained by Olson with reference to encompassing organizations and peak associations:

... Norway's main special interest organizations are highly encompassing, especially in comparison
those in Great Britain and the United States, and probably are more encompassing than
those in any other developed democracies (Olson 1982, 90).

Peak associations, similarly,'... should on average take a somewhat less parochial than the narrow associations of which they are composed ...' and such organizations are 'unusually important' in Scandinavia (Olson 1982, 50). At this point two comments are appropriate. First, encompassing organizations and peak associations have quite different organizational patterns. Norway has very few encompassing organizations. The union of farmers (59,000 members) and the union of small holders (9,000 members) should probably be seen as encompassing in the agricultural sector. These unions have been in sharp competition and conflict during recent years. Peak associations are, however, a rather common phenomenon. Among employers there is one dominating peak (NAF) with about 9,000 firms and 50 national organizations as members. The situation for workers and professionals is less transparent and more competitive. With a total of approximately 230 organizations, about 75% are knit to one of three dominating peak associations.¹ The organizational pattern of Norway may be less complex than the pattern in any other developed democracies, but nevertheless sufficiently complex to provide for competition, conflict and Chicken games - and even Prisoner's Dilemmas in some sectors - among organizations. then is our second comment: Given that some dilemma game still is relevant, it cannot be inferred that creative strategies are likely. Such inference:» are not possible without further information and additional reasoning. Hence, we have to move to institutional details.

Two institutional mechanisms now seem to be especially relevant, and both of
them induce organizations - both small and large - to act in a more social way
than they otherwise would do.

Interest Intermediation

In Norway, battles between interest groups are highly regulated. The anarchistic 'war of all against all' is replaced by corporate pluralism or a system of corporatist between organized groups and the state apparatus (cf. Rokkan 1966, Schmitter 1981). The system has even developed in the municipalities, a 'local corporatism' (Hernes & Selvik 1981). Interest intermediation takes many forms.

The most frequent example is participation in the networks of governmental committees, commissions,
and boards that initiate, design, advise and decide upon, implement and administer
policies (J.P. Olsen 1983, 148-49).

Norwegian corporatism is doubletracked; organizations obtain policy influenceat expense of restricted freedom of action. The latter fact, due to specific norms, rules and routines which guide the complex pattern of integrated organizational participation, is overly important from the governmental point of view. Collective decisions inside the 'thousand committees' where government authorities meet with organizational representatives of various conflicting and

overlapping interests and sectors of the society, are rarely transformed into governmental policies unless they are based on unanimity or compromise. Hence, it is better for interest groups to rule out strategies which do not contributeto advancement of compromises - at least where influence rather than symbolic protest is the goal. The system of interest intermediation leads us to suggest that only strategies which at least in some respect are creative are politicallyadmissible. and negotiations cannot be won by open egoistic - i.e. purely distributive - approaches, as interest groups will not be listened to nor taken notice of if they adopt this attitude in political processes such as legislation or planning. There will be a constant need to relate one's own behaviour to the interests of society as a whole or to principles of justice, equality, etc. Hence, incentives embedded in the system of interest intermediation point in the directionof choices in the dilemma games outlined above. The question of strategy choices in small and encompassing organizations thus cannot be settled on purely logical or deductive grounds; only empirical studies can help us settle the matter. This conclusion can be reinforced by looking at the phenomenon of cost externalization - underlying Olson's theory - in the Norwegian context.

Representation, Decision Rules, and Cost Externalization

Cost externalization has taken place when benefits from political decisions are concentrated to particular interest groups, and costs are dispersed throughout society. Net benefits to an interest group, for example in the wake of some lobbying may be very high even in cases where the rest of the society finds the costs to be infinitesimal. The essence of the phenomenon is depicted in Table 1. If project I is adopted, interest group 1 gets high net benefits and the rest of the groups have to share the expenses. If enacted, we assume that each of the projects inhibits economic growth.

How is it possible to get any of the projects adopted? If the projects are voted
upon one by one, and decided by majority rule, none of them will survive. Thus,

cost externalization seems to require some kind of 'minorities rule*. One possibilitythen of platform voting; by logrolling, interest groups can form winning coalitions to get several projects adopted. Another possibility is to establish a system of committees; with only a few interest groups represented in each committeee, all of the projects may pass. For example, project I, II and 111 are decided upon in a committee where groups 1,2 and 3 are represented.

Above, unanimity was mentioned as a central - or even the dominant - decision of Norwegian corporatism. Furthermore, quite a few - if not the greater part - of the boards, councils, committees and commissions are characterized by rather broad interest representation. In combination, these two institutional aspects probably make the phenomenon of cost externalization less important in Norway than in several other Western nations. And in our opinion, thus rendering theory less convincing or applicable in the case of Norway.

Economic Growth: A Public Good?

In his introductory chapter, Olson (1982, 18) informs us that his new book is an outgrowth of The Logic of Collective Action (Olson 1965), partly'... an application the argument in it...'. Furthermore,'... the most serious critics or students of the present book should have read that one' (Olson 1982, 18).

Olson's propositions are derived from a relatively simple and formalized theory which is built on explicit assumptions. He posits a particular mechanism through which distributional coalitions have impact on various variables, including growth. This mechanism is called The Logic of Collective Action'.

The theoretical model is founded upon fairly restrictive assumptions. In particular, should like to focus on the nature of collective goods. A public good is defined '... as any good such that, if any person Xj in a group Xj..Xj..X_n consumes it cannot feasibly be withheld from others in that group' (Olson 1965, 14).

We may notice that this definition is somewhat different from Samuelson's concept of public goods. The Olsonian collective good is not joint in the sense that making it available to one actor means that it can freely be supplied to other actors. Olson speaks of 'fraction of the group gain', an expression which does not apply unless the good can be partitioned into pieces which are consumed individually.

Now, we may ask whether the theory of Olson may be applied in the analysis of economic growth. Olson suggests that'... members of the organization would get only a part of the benefits that would result if they made the society as a whole more efficient' (Olson 1982, 4). Is it obvious that members of organizations obtain minor fractions of the increased efficiency? Can economic growth be assumed to be a non-excludable good?

Snidal (1979) argues that control over exclusion to a large extent is dependent on social institutions. These vary considerably between various politicoeconomic The major instrument for the redistribution of economic values is central government. At one end of the scale, government plays a minor role in the economy, leaving problems of production and distribution to the market, to the firms and the employees. At the other end, economic values are collected as taxes being redistributed independent of who produced the economic in the first place. This is the case of a welfare state. We propose that to the extent government provides equal rights to consume the 'social pie', the assumption of economic growth being a public good seems to be tenable. Thus, the size of the public sector - the extensiveness of the welfare state - is a major determinant of the publicness or collectiveness of economic growth. Therefore, we may apply the theory of Olson to explain economic growth in countries with a large public sector, while the reasoning seems less applicable in countries with a small public sector.

However, it is arguable that economic growth is a public good because it indirectly everybody in society. Surely, the value of individual wealth may be dependent on the general economic situation. Positive spillovers generated by economic growth may increase the overall standard of living. Even so, it seems quite clear that such a wide concept of public goods does not coincide with Olson's usage of the term. Positive externalities do not necessarily remove incentives increase productivity. To the extent increased efficiency is individually profitable, each will contribute whether or not spillovers are generated.

As indicated previously, Olson argues that a system of comprehensive interest groups provides some motivation to increase the social pie. However, it is arguable whether comprehensive organizations do in fact contribute more to economic growth than do smaller organizations. We shall attempt to analyze this hypothesis by a simple mathematical model. The following model represents only one possible representation of the Olsonian argument. Although this is the case, we feel that our formalization is quite close to the original model.

Let us first define two crucial variables:

C : Total amount of collective good, i.e. growth of GNP

Q: The fraction of collective good available to actor i

This gives a defining equation:

(1)

Let us then suggest a simple micro-production function:

(2)

Here, the variables are interpreted as:

gj: Amount of collective good (GNP growth) provided by actor i

eji Effort by actor i to increase GNP growth or collective good

In order to establish the macro-production function, it is assumed that aggregate is a weighted sum of micro-production functions. The weights represent the size of the actors. Aggregate GNP-growth (amount of collective good) may thus be written:

(3)

Amount of public good

Lastly, we should model the mechanism through which the collective good of economic growth is distributed. Provided that growth is a pure private good, each actor will get what he produces himself. Under such a free-capitalist system, it is reasonable to assume a maximizing behaviour in the following sense:

(4)

'Capitalist model'

In a pure welfare-state regime, each will have equal access to the collective
good; that is Cx =CC2= . .Cj=Cn. Now, the actor may maximize

(5)

'Welfare state model'

Neither of the polar cases give a quite good representation of Olson's model In The Rise and Decline of Nations, the actors may employ resources to get larger shares of the social pies. We should therefore endogenize the variable a^and suggest that distributional coalitions may employ political resources to increase their share of the economic surplus:

(6)

Here, Z, measures the political effort of actor i to increase his own part of the
pie - the collective good of GNP growth.

Let us assume that each actor is taxed proportionally, and that central government redistributes the whole social pie as cash transfers. Furthermore, that each actor obtains some gain from his own production, and some gain from lobbying government in order to get slices of the public pie. The model may then be formulated as in (7):

(7)

when

'Mixed welfare state/capitalist model'

In this model the tax rate is X. Here, economic growth generates a mix of
private and public benefits. Accordingly, the actors maximize the net benefits
from both private and public goods.

In order to simplify the argument, we assume that the social pie is
redistributed through central government, meaning that the tax rate is X = 1.
This gives a somewhat simpler model:

(8)

'Olson model'

Here, it is assumed to be no upper restriction on the use of resources. This is
reasonable, provided that resources put into growth-increasing policies (q) do
not exhaust resources increasing the share of the pie (Z;). Maximizing (8) yields

(9)

(9) defines the solution (q*, Z*). Comparing with the capitalist model (4), it is clear that fewer resources are put into production activity. (The capitalist solution is f{ (q) = 1.) Now, we may ask whether a system of encompassing organizations produces a higher rate of economic growth. Within the model, the distribution of B; indicates the size-distribution of the actors. Moreover, an increase of fy in the equilibrium (9) produces a higher value q. (f" (q) < 0). Also, Zj will increase.

Therefore, encompassing organizations may both employ larger resources to get larger slices of the public pie as well as higher efforts to increase this pie. This point seems to be somewhat obscured in Olson's book (p. 74, 80). The above result is, however, sensitive to modifications in the model. If a fixed amount of resources is assumed (q + Z; = r; (rjfixed)), the question whether q or Z; will increase seems to be inconclusive. Hence comprehensive organizations may divert larger resources to increase the collective good of economic growth and smaller amounts to obtain slices of the good. But the effects may also go in the opposite direction.

As far as Norway is concerned, the above reasoning casts some doubt on the
hypothesis that the relatively high economic growth rate is due to the

encompassiveness of Norwegian organizations. The theoretical argument for
this hypothesis may need further clarification.

Conclusion

This article has argued that reality is more complicated than Olson's elegant theory leads us to believe. This does, however, not necessarily imply a refutation of the theory. Our task has only been to throw doubt on the implications concerning group behaviour and economic growth. Are the assertions of The Rise and Decline of Nations justifiable, or are the arguments lacking essential details?

---

NOTE

1. Data reported in this section are mainly from NOU 1982, No. 3. See also Heidar 1983.

REFERENCES

Heidar, K. 1983. Norske Politiske Fakta 1884-1982. Oslo: Universitetsforlaget.

Hernes, G. & Selvik, A. 1981. 'Local Corporatism', in Berger, S. ed. Organizing Interests in Western
Europe. Cambridge: Cambridge University Press.

NOU 1982, No. 3. Maktutredningen. Sluttrapport. Oslo: Universitetsforlaget.

Olsen, J.P. 1983. Organized Democracy. Political Institutions in a Welfare State - The Case of Norway.

Olson, M. 1971. The Logic of Collective Action. Cambridge: Harvard University Press.

Rokkan, S. 1966. 'Norway: Numerical Democracy and Corporate Pluralism', in Dahl, R.A. (ed.): Political
in Western Democracies. New Haven: Yale University Press.

Samuelson, P.A. 1954. 'The Pure Theory of Public Expenditures', The Review of Economics and
Statistics 36, 387-389.

Schmitter, P.C. 1981. 'Interest Intermediation and Regime Governability in Contemporary Western
Europe and North America', in Berger, S. ed. Organizing Interests in Western Europe. Cambridge:
University Press.

Snidal, D. 1979. 'Public Goods, Property Rights, and Political Organizations', International Studies
Quarterly 23, 532-566.