Testing the Implications of the Olson Hypothesis

Considerable attention has been paid to the hypothesis, advanced by Olson, that accumulations of interest groups build up over time and retard economic performance. Most previous tests of this hypothesis have been based on comparisons of growth rates. In this paper, it is argued that comparisons of income levels yield a superior test. Strong forms of the Olson hypothesis, in which shocks such as revolution and war yield positive benefits are rejected using this test. I would like to thank Mancur Olson for helpful comments and criticism. Testing the Implications of the Olson Hypothesis Since about 1970, economic performance throughout the industrialized world has been significantly worse than in the earlier post-war period. Also over the post-war period, there have been significant differences in performance between countries, with, for example the US and UK growing relatively slowly and Japan and West Germany growing relatively rapidly. Various models have been advanced to explain these phenomena, but there is obvious interest in a hypothesis that could explain both of them. Olson (1982, 1988) advances such a hypothesis, based on the notion that accumulations of interest groups build up over time and retard economic performance. Shocks to the social order, such as revolution or defeat in war, may break up these accumulations and therefore promote more rapid growth. In addition to explaining relative economic growth and macroeconomic performance in the post-war period, the theory is applied to the larger question of The Rise and Decline of Nations. Most tests of the Olson hypothesis, including those reported by Olson (1982) and those in the volume by Mueller (1983) have involved comparisons of growth rates, particularly between nations that have experienced shocks and those that have not. In this paper, it is argued that these tests have failed to compare the Olson hypothesis with a natural null hypothesis, in which there are no benefits to the dissolution of interest groups. A preferable test of the hypothesis involves comparisons of levels of income and welfare rather than growth rates. It is also necessary to re-examine the implications of the hypothesis for the policy process. On the face of it, the hypothesis appears to suggest that a determined assault on 1 interest group power is a necessary precondition for a return to dynamic economic growth. Given the examples of Japan and West Germany, which achieved rapid growth after the devastation associated with World War II, it would seem arguable that such an assault would be beneficial even if the transitional costs were very high. In this paper, it is shown that the strong versions of the Olson hypothesis, suggesting that destructive shocks may have positive net benefits, cannot be sustained by a valid empirical test. It follows that transitional costs must be taken into account in any policy program aimed at reducing the influence of interest groups. I. OUTLINE OF THE HYPOTHESIS The central idea in Olson’s work is that economic performance is hampered by the activities of interest groups. In itself, this idea is scarcely new it has been a staple of political rhetoric for millennia. What is novel in Olson’s presentation is the derivation of a model of interest group activity based on his (1965) Logic of Collective Action, and the use of this model to derive predictions on the comparative impacts of interest group activities across societies and time-periods. The most important new idea is that, in a stable democratic system, interest groups tend to accumulate over time. The underlying reasoning is that interest groups are costly to organize, but that, once organized, such groups are unlikely to dissolve even if their support base shrinks or their original rationale is lost. A corollary of this analysis is that shocks to the system, which break up coalitions can have beneficial long-term effects. The crucial positive illustrations are the ‘economic miracles’ experienced by Japan and the Federal Republic of Germany after World War II. The negative illustrations are the low growth rates experienced by the Anglo-Saxon countries, which have not experienced shocks. The most notable case is the relative economic decline experienced 2 by the United Kingdom, which has gone from being the world’s richest and most powerful nation in the 19th Century to a position in the lower half of the OECD ‘league table ’ of per capita income today. There is a natural analogy with the debate over the role of depressions in a market economy. The orthodox Keynesian view is that depressions are disastrous events, and that the prime object of economic policy should be to prevent or mitigate them. This view has been contested by a number of writers, most notably Schumpeter (1961). Schumpeter argued that the ‘creative destruction’ inherent in depressions and crises weeds out inefficient and bureaucratized firms, and paves the way for a new surge of entrepreneurial vigor. A similar view may be found in Marx (1867), although he saw the crisis as part of a dialectical process that is both necessary to the operation of capitalism and destined to lead to its collapse. The Olson hypothesis may be seen as a translation of Schumpeter’s view from the economic to the political sphere, with political shocks taking the place of market fluctuations. In both models, the central notion is that the competitive market economy is the engine of growth. The liberal democratic state is seen as the natural counterpart of the market economy. Both contain the seeds of their own destruction. Schumpeter lays more stress on the realm of ideas. He sees industrial capitalism as calling forth an intellectual class that is naturally inclined to a rationalist, rather than en entrepreneurial ethic, and that is ultimately supportive of socialism. By contrast, in common with the dominant trends in public choice theory, summarized by Mueller (1979), Olson (1982) pays little attention to the rôle of ideas, and focuses on interest groups (see however, the discussion of this issue in Olson 1989). Olson’s model is based on the assumption that groups rationally pursue their own interest through political 3 1 There is a clear contrast with the real business cycle theory in which crises are treated as optimal equilibrating responses to external shocks. There is no room for entrepreneurship in a model of this kind. 2 Depressions are not agents of creative destruction in Olson’s model of the political process. Indeed, Olson (1989) argues that depressions tend to call forth increased activity by narrow distributional coalitions seeking to resist income losses. processes. The pluralistic democratic process rewards well-organized interest groups and results in their proliferation. This retards the competitive engines of growth and leads to increasing political division and, ultimately, an ‘ungovernable’ society. A second aspect of Olson’s analysis, which has received much less attention than the notion of ‘creative destruction’ is the idea that interest groups may be benign if they are ‘encompassing’. An encompassing interest group is one whose membership is large relative to the economy as a whole, such as a national confederation of unions or employers or a broadly-based political party. Because of their size, these interest groups cannot pursue redistributional policies without regard for the effect on the total size of the cake to be divided. Hence they will have an incentive to seek efficient methods of redistribution and to avoid proposals that greatly reduce aggregate income. The test proposed here may be applied to examine the impact of encompassing interest groups on long-run performance. On a related theme, Olson argues that ‘jurisdictional integration’, arising, for example, from the federation of previously independent states, may yield benefits because interest groups operating at the level of the individual states will become less effective and interest groups at the federal level will take time to form. Jurisdictional integration is less costly than the more radical shocks mentioned above, and is therefore more likely to yield positive net benefits. Like Schumpeter, Olson does not put his reasoning in strictly formal terms. As a result, it is not immediately clear how his hypothesis should be tested. It is necessary, therefore to present Olson’s hypothesis in the context of a formal model.