Model uncertainty, political learning, and institutions: a broader view on Mancur Olson's theory of institutional sclerosis.

1. Introduction

Twenty-five years ago, Mancur Olson published his monograph on the Rise and Decline of Nations. Therein he argues that countries that have been politically stable for a long time suffer from institutional sclerosis. This sclerosis is caused by the growing political influence of so-called distributional coalitions. Their search for rents leads to increasing complexity in regulation and larger governments and reduces society's ability to adjust to changes in technologies and the economic environment. In the long run, declining rates of economic growth are the result. Olson's Rise and Decline of Nations marked a milestone in the research on the political economy of growth and inspired a large body of literature. In particular, the fact that he provided a number of testable hypotheses invited many scientists for empirical testing (among others, Choi 1983; Murrell 1984; McCallum and Blais 1987; Gray and Lowery 1988; Tang and Hedley 1998; Bischoff 2003) and still invites more testing (Mueller 2003). The existing empirical evidence is mixed and provides only limited support for Olson's theory.

This paper takes this mixed empirical evidence as its starting point. Section 2 briefly reviews the basic chain of cause and effect underlying Olson's theory and categorizes the related empirical literature with respect to the different elements of this chain. Section 3 sketches a simple model of party competition in a complex environment, and section 4 shows how this model can be used to provide an alternative explanation for low growth rates in politically stable democracies. In the light of these considerations, section 5 reassesses some of the anecdotal evidence Olson provided to motivate his theory and the mixed empirical support for his theory. It also draws on new empirical evidence on the relationship between institutions and economic growth and shows how this literature relates to Olson's theory, to the theoretical model sketched in section 3, and to the empirical literature reviewed in section 2. It will provide additional reasons why the support for Olson's theory is likely to be mixed and points out directions for future research. Section 6 concludes.

2. Olson's Rise and Decline of Nations--A Brief Outline

The Rise and Decline of Nations does not provide a full-fledged formal model to back its central theme. Instead, it relies on economic intuition, a rich body of case-based evidence, and a number of regressions on the role of unions and the relationship between years of statehood and growth in the United States. Following Gray and Lowery (1988), Unger and van Waarden (1999), and Paque (1996), the basic chain of cause and effect underlying Olson's theory consists of three hypotheses (see Table 1). If all three of these are valid, the chain holds, and thus Olson's theory provides a valid explanation for declining growth rates in politically stable countries. If one element within the chain does not hold, it does not necessarily mean that politically stable countries are not suffering from institutional sclerosis and declining growth rates. In this case, however, Olson's theory does not provide the correct explanation, but other explanations must be found.

When analyzing and/or testing Olson's theory, it is possible to take steps of different scope. The right-hand side of Table 1 shows that, depending on the scope of the analysis, five different types of studies can be identified. First, it is possible to take on a single hypothesis and assess its tenability individually (see studies of type I, II, and Ill in Table 1). The literature contains a number of empirical studies of type I, testing the relationship between the number of interest groups within a country and its history of political stability. The evidence is mixed and proves very sensitive to the definition of political stability (Murrell 1984; Bischoff 2003: Coates, Heckelman, and Wilson 2006). Second, one can take two hypotheses of Olson's chain of reasoning and test them jointly. A number of studies performed a joint test of hypotheses 2 and 3 to see whether the number of interest groups has a negative effect on economic growth (Gray and Lowery 1988; Knack 2003). The number of these type IV studies is limited because of restrictions in data on the number of interest groups. The majority of studies compiled by Heckelman (2007) found weak or no support for a negative relationship between the number of interest groups and economic growth. Third, by far the largest part of the literature tested all three hypotheses jointly by analyzing the direct relationship between a country's years under political stability and its economic growth rates (Choi 1983; McCallum and Blais 1987: Tang and Hedley 1998). The majority of studies find a negative relationship, but other results are also reported. Mueller (2003) and Heckelman (2007) conclude that the existing evidence generally confirms Olson's central conclusion. It is important to note that most empirical studies are based on cross-sectional data. They show that countries with a long history of political stability have lower growth rates than countries with a shorter history. This result suggests that the economic growth rates of a single country decline with the duration of its political stability. Nevertheless, when referring to the empirical evidence, I will hereafter speak of low growth rates in politically stable countries (rather than declining growth rates).

Olson's theory was discussed among economists, political scientists, and historians (Heckelman and Coates 2003). In many cases, they challenged not his central conclusion but rather the underlying chain of reasoning (Maddison 1988). Some authors, particularly historians, argued that Olson does not account for the many country-specific factors that determine differences in economic performance (Paque 1996: Whaples 2003). Among economists, the neoclassical catching-up hypothesis is considered an important alternative explanation for the empirical observation that politically stable countries grow at a slower pace than countries that recently endured politically unstable times. Especially after World War II, countries such as Germany and France, whose physical capital stock had been largely destroyed by the war, naturally experienced high rates of capital accumulation and thus growth. They were catching up with Great Britain and other countries that were less heavily destroyed after the war. If duration of political stability is generally positively correlated to GDP per capita, catching up provides an alternative explanation for low economic growth rates in politically stable countries (Abramovitz 1983; Nardinelli, Wallace, and Warner 1987; Maddison 1988). Accordingly, Olson's empirical observation may be caused not by the negative impact of political stability on growth but rather by political instability frequently coinciding with a destruction of capital stock. Other authors point at an argument that has already been made by Olson (1982), but that is very difficult to account for in empirical studies. Accordingly, the third hypothesis of Olson's reasoning holds only for special interest groups but not for encompassing groups. Societies with strong encompassing groups are less likely to witness declining growth rates than countries that are populated primarily by special interest groups (Maitland 1985; Gray and Lowery 1988; Lindbeck 1997). This may explain why the evidence of type V and especially type IV studies is mixed.

Sections 3 and 4 provide an explanation for this empirical picture, which applies especially to democracies and that has so far been neglected. The starting point is a simple model of party competition in a complex environment where voters and politicians do not perfectly understand the functioning of the economy and of economic policy. I will argue that party competition is a path-dependent process and that politically stable countries will find themselves in a poorer starting position when having to adjust to changes in the economic environment. Other things equal, they will fail to adjust more frequently and thus witness periods of low economic growth due to inadequate policies. The model does not require an active part of distributional coalitions and thus may explain why the evidence for type IV studies is mixed. Section 5 leaves the narrow focus of the model. It draws on more recent literature on institutions, political stability, and economic growth. I will argue that some institutions are able to increase the probability of successful adjustment in politically stable countries. Similarly, institutional settings codetermine the political influence of distributional coalitions. In both cases, the degree of institutional sclerosis depends on the political institutions. The fact that this has rarely been accounted for in the empirical studies of type IV and V may explain their mixed results.

3. Political Decision Making and Party Competition under Model Uncertainty

The section starts by assuming that the economy is a complex system. Its macroeconomic performance is determined by a large number of different factors, some of which are unobservable, and the relationship between these different factors is nontrivial (von Hayek 1972; Rosser 1999). Because of cognitive limitations, neither voters nor politicians are able to gain a full understanding of the economy and the impact of different economic policy measures. In other words, they act under model uncertainty. Under model uncertainty, adequate policies cannot be identified ex ante in thought experiments but must be discovered in real experiments (von Hayek 1972; Wohlgemuth 2002). In parliamentary democracies, the process of choosing the experiments to be conducted consists of two steps. First, the political parties suggest certain experiments (by proposing policy platforms). Second, the voters choose between proposed experiments (by voting for a particular party). The electorate's approval is necessary before any experiment is conducted.

Given the minute effect of the individual vote on the electoral outcome, voting is primarily an expressive act motivated and guided by emotions, beliefs, and easily accessible information (Mueller 2003). Rather than efficiency, fairness and distributional concerns provide the adequate yardstick for this kind of judgment (Clark 1998; Fechtenhauer et al. 2006; Haferkamp et al. 2006). Other things equal, a party can expect more votes the more its policy platform appeals to voters on distributional and fairness grounds. On the other hand, the empirical research on vote-popularity functions shows that voters hold the government responsible for the macroeconomic performance, in particular unemployment rates and GDP growth. They punish governments for bad economic results and reward them for good performances even if their individual prospects are not affected (Paldam 2004). In public choice models, the impact of economic performance on voter behavior is captured by the concept of valence (Ansolabehere and Snyder 2000; Groseclose 2001). A party's valence reflects the degree to which voters consider this party capable of producing good economic outcomes. The valence of a party is a function of the economic results this party achieved during previous terms in office (Norpoth 1996; Paldam 2004). Other things equal, a party that has produced better macroeconomic outcomes in the past is attributed a higher valence and can thus expect more votes.

A large number of policy fields (including public infrastructure, taxation, public education, foreign trade policies, social security, environmental policy, and regulation) are recognized to influence macroeconomic outcomes. In any of these fields, numerous different policy instruments can be applied in different combinations and intensities. In other words, there are a large number of policies that could be pursued. In the context of this paper, two attributes can be used to characterize policy platforms. Some policies are popular, meaning that they appeal to a large number of voters on distributional and fairness grounds. Growth-promoting policies yield high growth rates when pursued. A large number of policy platforms are likely to be neither popular nor growth promoting.

Following the public choice logic, political parties can be expected to search for policy platforms that would generate the majority of votes in the current as well as in upcoming elections (Enelow and Hinich 1984). In the short-term perspective, popular platforms are very attractive for political parties because they dominate all other policy platforms as long as parties are assigned the same valence. A popular policy platform can be defeated in an election against an unpopular policy platform only if the party offering the latter is attributed a higher valence. The disadvantage of offering popular policy platforms is that they may yield poor economic results and thereby reduce the chance to win future elections. The time horizon of most parties exceeds one election term. Therefore, parties may prefer growth-promoting policy platforms because these bring a high valence and increase the number of expected votes in future elections. (1)

In a complex environment, the parties initially do not know which policy platforms are growth promoting and which ones are popular. This information is gradually revealed in consecutive elections and government terms. In the beginning of this process, party members have convictions as to which policy platforms are popular and/or growth promoting. These policy platforms constitute the different parties' knowledge pools. Different parties are likely to start with different knowledge pools because their members differ with respect to social background and convictions (Denzau and North 1994; Tetlock 1999). Each election generates some feedback about the popularity and economic success of the policy platforms offered in the elections and pursued thereafter. Based on this feedback, the parties learn. Learning in this context is equivalent to changing the knowledge pools by adding new platforms and dropping old ones (Edmonds 1999). As the stakes are high, politicians can be assumed to spend substantial effort in the search for policy platforms that yield high numbers of expected votes for current and future elections. They may turn to political advisers or draw on evidence from other countries. Voters face negligible returns from learning about economically superior policy platforms only because they cannot expect any additional utility from making a better-informed choice (Mueller 2003). Thus, the electorate as a whole will learn only very slowly from previous elections, policy advice, or evidence from other countries. In other words, the distribution of voters' preferences is stable in the short- and medium-term perspective. Radical shifts can be expected only in the case of severe crises, wars, or massive migration.

Given this starting point, the essential question is whether party competition will lead to the discovery and application of growth-promoting policies. Two cases can be differentiated. If there are policy platforms that are both popular and growth promoting, there is no trade-off between fairness and distributional concerns on the one hand and macroeconomic requirements on the other. Parties will be led to discover good economic policies just by following the interests of the voters. By maximizing votes in the short-term perspective, they increase their valence and thereby increase their chance of winning elections in the long run. Successful learning, that is, the discovery of growth-promoting policies, is self-evident.

Frequently, the electorate's ex ante preferences with respect to fairness and distributional concerns will stand in contrast to the macroeconomic requirements (Persson and Tabellini 2006; Sunde 2006). In this case, policy platforms that are popular ex ante will yield poor economic growth rates ex post and vice versa. From the parties' point of view, there is a tradeoff between maximizing votes in the current election and maximizing the chance of winning future elections. This trade-off occurs because voters update their valence only ex post: A party can maximize its chance of winning the current election by offering a popular policy platform. At the same time, however, this step may reduce its valence and thus reduce expected votes in future elections. Offering a growth-promoting policy platform can be thought of as an investment. It reduces expected votes in the short run but raises the party's valence and thereby increases the chance of winning elections in the future. However, this investment pays only if the party manages to win the current election. This precondition is met if the political opponents offer nonpopular policy platforms or if the investing party has an advantage in valence that outweighs the lower ex ante popularity of growth-promoting policies. This points at a crucial determinant of successful learning: the contents of the parties' knowledge pools in the early stage of the learning process. If the knowledge pools contain popular platforms, successful learning is unlikely because growth-promoting policy platforms have only a very slim chance to attract the majority of votes and prove their economic superiority. If, however, all parties start off with unpopular policy platforms, growth-promoting platforms can prevail and prove their superiority. Thus, the discovery of growth-promoting policies is possible but not self-evident.

If the process of learning does not lead the parties to discover growth-promoting policies, the knowledge pools will soon be dominated by popular policy platforms. In this case, parties offer policies that appeal to the voters on fairness and distributional grounds but yield poor economic results. No party can expect to win an election by switching to growth-promoting policy platforms because, given the parties' poor valence, these would not obtain voters' approval. On the contrary, the party would harm its own interest by giving away the chance to win the current election. This situation represents a stable political equilibrium with low economic growth rates (hereafter "low-growth equilibrium"). There are two ways to escape this equilibrium. First, major parties within a country may form a temporary cartel in which they agree to perform unpopular but potentially growth-promoting policies. This step can break up the strategic dominance of popular policies because the parties involved in the cartel benefit from the potential increase in valence in the long run without giving away the chance to be in office in the short run. In reality, such "grand coalitions" or "governments of national unity" are sometimes observed in countries that witness a deep and persistent crisis that was not solved by previous majority governments. For example, Germany currently witnesses its second grand coalition, partly to deal with the aftermath of the reunification. The parties explicitly stepped up to overcome the strategic deadlocks that result from the short-term political costs of growth-promoting reforms and to pursue policy platforms they consider necessary to guarantee economic growth in the future. The second way to escape the low-growth equilibrium consists in a change in the distribution of voters' preferences. North (1990) pointed out that preferences are likely to be endogenous in the long run and that one might expect a change in preferences if the current preferences prove an obstacle to economic development. The gradual shift toward accepting positive interest payments among peers in medieval Europe (Fuhrmann 1996) or the growing acceptance of market-oriented institutions in contemporary Eastern Europe can serve as examples. These examples indicate that countries caught in the low-growth equilibrium are not doomed to remain trapped forever but that changing preferences can help them escape it. However, given the slow pace at which preferences change endogenously, they cannot expect to escape the trap within only a few years but are likely to witness a longer period of stagnation.

4. Structural Shifts and the Problem of Low Growth Rates in Politically Stable Countries

This section returns to Olson's theory of institutional sclerosis. It applies the conclusions from section 3 to provide an alternative explanation for low growth rates in politically stable countries. Accordingly, low growth rates result from the notion that politically stable countries lose their ability to adapt to strong shifts in the structure of the economy, which require substantial changes in economic policy. These structural shifts may result from permanent shifts in relative prices, new technologies, vital resources becoming scarce, the breakdown of old markets and the discovery of new ones, the emergence of new competitors, or a change in international economic institutions. In the 20th century, the two world wars, the breakdown of colonialism, the end of the Bretton Woods system (Maddison 1988), the introduction of the European Common Market (and Currency), and the transformation of Eastern Europe and China constituted such structural shifts. A structural shift means that former growth-promoting policies no longer yield good macroeconomic results and that some of the policy platforms that produced poor results prior to the shift become growth promoting. Under model uncertainty, these new growth-promoting policies are unknown ex ante and have to be discovered in real experiments. The claim will be made that politically stable countries are less likely to discover these new policies because their starting position is less favorable. As a result, they are more likely to be caught in a low-growth equilibrium after a structural shift than countries that have recently witnessed political instabilities. (2)

To illustrate the basic logic, I follow Olson (1982) in conducting a thought experiment involving two countries, S (stable) and U (unstable). Country S has been politically stable for a large number of periods, while country U has witnessed a period of political instability. Now assume that both countries witness a structural shift. The crucial difference between country S and country U is given by their different knowledge pools at the time of the shift. The starting position of country U is greatly influenced by having endured politically unstable times. First, the political instability interrupts the process of learning. Second, political instabilities of the type that Olson had in mind are very disruptive to the distribution of voters' preferences. As a result, the political parties that struggle for power after the unstable period lack reliable information about the distribution of voters' preferences. This applies to the historical example of World War II, which Olson used to motivate his Rise and Decline of Nations. After democracy was (re)established in Germany and Japan, none of the political parties had reliable information about the voters' preferences. Information about voters' preferences prior to the war cannot help the parties identify popular policy platforms because the war massively changed both the pool of voters and their preferences. In other cases, the countries were not democratic prior to the time of political stability, and thus the parties cannot refer to older information about voters' preferences. This applies to the parties and movements struggling for power in young democracies like India after gaining independence or many Eastern European countries after the fall of the iron curtain. Third, the political movements and parties in countries that recently witnessed political instability are frequently very polarized in their political convictions (Lebas 2006). As they are often dominated by single personalities rather than by well-established parties, they are often volatile in their policy decisions. Finally, the number of movements tends to be large, and the political competition in the early rounds of the learning process will decide which ones prevail. In sum, there is a substantial degree of heterogeneity in the starting knowledge pools in country U, and it has a reasonable chance to discover growth-promoting policies.

Now consider the situation of country S. The fact that it has been politically stable for a considerable time means that the parties had time to differentiate popular and possibly also growth-promoting policy platforms from other platforms. If country S was caught in the low-growth equilibrium prior to the shock, popular policy platforms will dominate the parties' knowledge pools. If country S has been economically successful, at least one knowledge pool will contain policy platforms that were growth promoting prior to the structural shift. In addition, the parties' knowledge pools are likely to contain popular policy platforms because these remain attractive in a short-term perspective. Regardless of the country's previous economic success, policy platforms that are neither growth promoting nor popular will not be included. This also means that policy platforms that are growth promoting after the structural shift are unlikely to be contained in any party's knowledge pools. Therefore, the ex ante probability that these new growth-promoting policies are suggested for a real experiment in the first place is very small. Instead, the parties are likely to continue to propose policy platforms from their preshift knowledge pools. The formerly growth-promoting policy platforms no longer produce good economic results and are thus eliminated. Soon, the knowledge pools are dominated by popular policy platforms, and the country becomes caught in the low-growth equilibrium.

It could be argued that parties can learn from countries that have successfully adjusted to a similar shift. Thus, some party might propose policy platforms that are adapted to the structural shifts and that would yield good economic results if they were applied. This raises the following question: Can a country prevent the low-growth equilibrium if one of the parties offered a policy platform that is growth-promoting after the shift? The answer is affirmative if there is a chance that a party offering this new platform wins the majority of votes. Unfortunately, the other parties are very likely to offer popular platforms because these are particularly attractive for parties after the structural shift. In this case, the party offering the new growth-promoting platform can win the election only if it is attributed a high valence because it has been successful in times prior to the shift. After the shift occurred, this high valence is in danger of deteriorating very quickly because policy platforms that produced the party's high valence before no longer yield good economic results. In other words, parties that were attributed a high valence because of their economic performance prior to the shift are in danger of losing this advantage quickly. This means that the superiority of the new policy platform can be discovered only if the party switches to this new platform while its advantage in valence still exists. Once the advantage in valence has deteriorated, the party can no longer win the election against the popular opponent platforms. Consequently, the time slot for a successful adjustment is narrow and can be passed only if the first policy experiments involve the correct policy platforms, that is, policy platforms that produce good macroeconomic outcomes after the shift. If these are not discovered quickly, learning will guide parties to offer popular policy platforms, and the country becomes caught in the low-growth equilibrium. This does not necessarily mean that the country cannot escape this situation, but it most surely witnesses a period of poor economic results.

In sum, the ex ante probability that parties in politically stable countries consider policy platforms that yield high economic growth rates in the postshift era is very low. If such new platforms are considered, they have a very small probability of winning voters' approval and proving their economic superiority. In country U, both probabilities are higher. Thus, country S is less likely to adapt to structural shifts than country U. Essentially, the rationale behind this argument boils down to a fundamental truth about learning: Adequate solutions are more likely to be discovered if learning starts off with a diversified pool of ideas or strategies (Riechmann 2001). In country S, the process of learning has eliminated all policy platforms that were not popular or growth promoting prior to the shift, thus reducing the diversity in known policy platforms massively. In country U, the political instabilities have prevented the knowledge pools from converging to a narrow set of platforms and thereby preserved the diversity in policy platforms. After a structural shift when diversity in knowledge pools is essential for successful adjustment, politically stable countries find themselves in a poorer starting position and are more likely to be caught in a low-growth equilibrium.

5. Party Learning under Model Uncertainty, Institutions, and the Empirical Evidence on Olson's Theory of Institutional Sclerosis--A Reassessment 

This section returns to the starting point of this paper, namely, the mixed empirical support for the central implications of Olson's Rise and Decline of Nations, and attempts to explain why the evidence is mixed. The discussion is partly based on theoretical considerations outlined in the previous sections. In addition, it draws on more recent literature on the relationship between political stability, institutions, and economic growth. The main focus of the analysis rests on democratic regimes because the theory of party learning under model uncertainty applies only to these.

The previous sections outlined a simple model of party competition under model uncertainty and showed that it contains an alternative chain of cause and effect to explain the central theme of Olson's theory of institutional sclerosis: low growth rates in politically stable countries. Both theories argue that low growth rates in politically stable countries result from problems of collective action. The most important difference between the two is that Olson's theory of institutional sclerosis is based on the idea of distributional coalitions whose struggle for rents actively obstructs necessary changes. The explanation I put forth here does not claim active obstruction to be the cause of institutional sclerosis but argues that it follows from path dependency in democratic learning. This difference may contribute to explaining the mixed empirical evidence on Olson's theory. Studies testing the relationship between duration of political stability and economic growth were generally supportive of Olson's major conclusion. At the same time, studies directly analyzing the relationship between number of interest groups and economic growth rates find only limited evidence for a negative impact of interest groups on economic growth (see section 2). Combining the evidence of these types of studies suggests that while Olson was correct in claiming a negative relationship between political stability and economic growth, his theory was less tenable with respect to the underlying chain of cause and effect. Other mechanisms must explain the low growth rates in politically stable countries. The previously sketched theory points at one possible mechanism that may be at work in democratic countries.

The previously sketched model can also be applied to some of the anecdotal evidence that Olson provided to motivate his theory (Olson 1982). He pointed at the puzzling fact that, after World War II, the defeated nations of Germany and Japan managed to rapidly recover and maintained strong and persistent growth, while Great Britain--supposedly one of the "winners" of World War II--witnessed economic stagnation. In Olson's logic, the period of political instability weakened or destroyed old distributional coalitions in Germany and Japan and allowed for a new start after 1945. In Great Britain, the old distributional coalitions did not lose political power and prevented the necessary reforms. Paque (1996) challenged Olson's perspective by asking a simple question, Why the 1950s and not the 1920s? He argued that Germany was in a similar position after World War I as it was after World War II. In both cases, Germany was defeated in a devastating war with millions of deaths and lost territories and then reestablished with a new economic and political system. Following Olson's theory, the degree of institutional sclerosis should have been reduced by both wars, and Germany should have experienced a period of fast and persistent economic growth in the 1920s as well as the 1950s. In reality, however, Germany succeeded in the 1950s but failed in the 1920s. Paque (1996) pointed at a number of different factors, including the high reparation payments, the 1929 crash, and the defects of the Weimar constitution, to explain the difference. The previously sketched model points at another possible explanation. One of the essential features of the West German restart after 1945 was the strong influence the Allied forces had on the economic and political constitution. They guaranteed that West Germany became a market economy with restricted government competences (Lampert 1997); private property and economic freedom were given constitutional status. In retrospect, most Germans agree that this facilitated economic success after World War II and would be unwilling to change it. In 1948, when the constitution was passed and the first general election was held, policies such as the nationalization of major industries, strong government controls over prices and wages, and heavy protectionism were popular among German voters (Lampert 1997). The fact that Germany's first real policy experiments were market oriented and thus the growth-promoting character of these policies was discovered must be attributed to the Allies' influence. Given the economic success of this policy, Germany witnessed a gradual change in voters' preferences in favor of the market economy. In the 1920s, the German policy experiments did not lead to comparable success, and the majority quickly turned to political parties that offered nationalization and government control over prices. Together with other factors, unfavorable political learning determined the poor economic results in Germany in the 1920s.

Starting in the 1990s, a growing number of authors have addressed the role of institutions in promoting or hindering economic growth (North 1990; Knack and Keefer 1995; de Haan, Lundstroem, and Sturm 2006). In his distinguished lecture on economics in government, Olson himself pointed at the importance of institutions (Olson 1996). Some authors focus primarily on the impact of institutions that directly affect private economic activity, direct investment, and capital accumulation. They show that market-oriented institutions, economic freedom, and the rule of law are growth promoting (de Haan, Lundstroem, and Sturm 2006; Feld and Voigt 2006). Other authors address the impact of political institutions, in particular the impact of democracy on economic growth (Przeworski and Limongi 1993; Keefer 2004; Persson and Tabellini 2006). They arrive at the conclusion that democratic societies tend to have larger governments and more redistribution, which reduces growth. On the other hand, democracy increases economic growth via a number of indirect channels. The net effect is generally found to be moderately positive (Butkiewicz and Yanikkaya 2006; Sunde 2006). Persson and Tabellini (2006) argue that "the devil is in the details" and thus that a more differentiated look at different democratic institutions is necessary to disentangle the different effects. Most of the type IV or type V studies that explicitly focused on Olson's Rise and Decline of Nations do not introduce control variables to account for the impact of economic freedom, rule of law, or democracy, let alone differentiate between different democratic institutions. On the other hand, only very few studies that focused on these latter variables accounted for Olson's explanation. The study by Feld and Voigt (2006) is an exception in this respect. Their primary focus rested on the impact of judicial independence on growth, but they also included the variable "age of the constitution" in their cross-country regression. Both judicial independence and "age of the constitution" are found to have a positive impact on economic growth. The positive impact of "age of the constitution" is not in line with Olson's theory or with the previously sketched model. The result suggests that judicial independence spurs private investment and thereby offsets some of the negative effects of political stability. With respect to Olson's theory, this result also suggests that judicial independence may weaken the position of distributional coalitions and their ability to extract rents from other parts of society.

An institutional setting that deserves special attention in the context of this paper is the one of federalism. In a federalist state, important decisions that affect economic growth are made at the regional or local level. The relationship between decentralization and economic growth is not monotonic, but empirical evidence suggests that the degree of decentralization within a country codetermines its economic performance (Weingast 1995; Oates 1999; Breuss and Eller 2003). In the context of model uncertainty and party learning, federalism serves as a laboratory in which numerous policy experiments are made at the same time (yon Hayek 1945; Strumpf 2002). This speeds up the process of learning and increases the probability that growth-promoting policy platforms are discovered somewhere in the federation. Once such policies are discovered, the voters in a poorly performing region have the possibility to learn that superior results are feasible if another policy is applied. The opposition in this region can point at the superior performance in the neighboring region and propose similar policies. As voters in one region can easily observe the superior performance in another region, they may be more willing to adopt policies from that region even if their opposition party is not ex ante attributed a higher valence. This reduces the attractiveness of popular policy platforms compared to a centralist state, in which feedback from neighboring regions is not so easily available. When confronted with a structural shift, federalist countries are more likely to adapt successfully because they test a larger variety of policy platforms than a centralist country. Regions that tested inadequate policy platforms at first can learn from more successful regions. Thus, federalist countries are less likely to get caught in the low-growth equilibrium. The empirical studies of type V, which test the relationship between political stability and growth, do not account for the impact of federalism. This shortcoming may explain why type V studies produce mixed results. (3)

With respect to Olson's theory of institutional sclerosis, federalism has two additional effects. First, a federalist country can expect to see a larger number of interest groups because there are many more avenues for rent-seeking activities than in a centralist state, other things equal. Second, decentralization induces competition among regions, provides voters with superior information, and introduces yardstick competition (Weingast 1995; Inman and Rubinfeld 1997; Oates 1999). Thereby, federalism reduces the possibilities of governments to redirect rents. In sum, federalist countries will have more interest groups, but their influence, other things equal, is smaller than in centralist countries. The type IV studies, which test the relationship between number of interest groups and economic growth, do not account for this possibility. This shortcoming can provide another explanation for the mixed evidence of these studies, especially in those cases where the analyzed sample contains centralist countries alongside with highly decentralized ones (Mueller 2003).

So far, the analysis has focussed on the impact of political (in)stability on economic growth. At this stage, it is helpful to take a brief look at the factors that cause political instabilities. In some cases, they are caused by factors from outside the affected countries. This is particularly true for the many countries that are invaded and occupied during wars (e.g., the Netherlands, Poland, Belgium, or Korea during World War II). To some extent, this applies also to former colonies whose struggle for independence was nourished by World War II and its aftermaths. In other cases, political instability is caused by factors from within the countries. There is some empirical evidence that it is the poor economic performance of the old system and its inability to reenter the path of economic growth that causes political instability (Feng 1997; Posner 1997; Sanhueza 1998). The new democracies emerging from the fall of the iron curtain are good examples in this respect. If poor economic results, especially in countries with low GDP, cause political instabilities, it is possible to conclude that countries can prevent this type of instability if they manage to remain economically successful. In order to remain economically successful in the long-term perspective, the political institutions must promote successful learning and adjustment to shifts and restrict the political influence of distributional coalitions.

These considerations point toward the following directions for future research. First, future empirical research to test for the relationship between political stability and economic growth must account for the so-far-neglected role of institutions. Such studies would bring additional insight into the relationship between political stability and economic growth and enhance our understanding of why the empirical evidence from tests of Olson's theory is mixed. Second, the analysis should focus more on the role of economic reforms. The theory sketched here suggests that it is essentially the ability to adapt to structural shifts in the economic environment that determines whether a country will witness long-term economic growth. Olson pointed in a similar direction in his implication number 7 from Rise and Decline of Nations. Therein he argued, "Distributional coalitions slow down the society's capacity to adopt new technologies and to reallocate resources in response to changing conditions, and thereby reduce the rate of economic growth" (Olson 1982, p. 65). It is widely recognized that autocratic and democratic regimes differ in their ability to adjust to changes. In the context of this paper, it seems promising to differentiate between different types of democratic regimes and analyze their impact on the societies' ability for successful reforms. The topic of economic reforms and the obstacles retarding or hindering them has been addressed in a number of articles (Rodrik 1996). This literature poses a good starting point to help answer the question of why democratic societies differ in their ability to undergo necessary reforms and to what extent these differences are caused by differences in political institutions.

6. Conclusion

In his Rise and Decline of Nations, Olson argued that politically stable countries will become victim of the distributional coalitions that inhabit them. Their number and influence will increase over time, and their lobbying for preferential policy will slow down technological progress and lead to inefficiencies. Thus, politically stable countries will witness declining growth rates as they remain politically stable. Olson used the term "institutional sclerosis" to describe the situation for these countries. From the very beginning, Olson's book has inspired many theoretical articles and empirical studies in the field of economics, history, and political science. The empirical research by and large supported the central conclusion of his theory but found only limited support for the notion that distributional coalitions and in particular special interest groups cause institutional sclerosis.

This paper takes this mixed evidence as a starting point. It outlines a simple model of party competition under model uncertainty, which shows that the discovery of growth-promoting policies does not automatically follow from voters rewarding good macroeconomic performance ex post. Based on these results, I argue that institutional sclerosis may result from political learning being path dependent. When confronted with the necessity to adapt to a shift in economic structure, countries with a long history of political stability find themselves less well equipped than other countries. Thus, politically stable countries are more likely to witness periods of poor economic growth. Just like Olson's theory, the theory here suggests that low growth rates in politically stable countries result from problems of collective action. The essential difference is that this new theory does not require distributional coalitions to actively obstruct growth-promoting reforms, which is central to Olson's theory. The two theories should not be viewed as mutually exclusive alternatives. Instead, both mechanisms may contribute to institutional sclerosis and low growth rates in politically stable countries.

In the past 15 years, a growing body of literature has investigated the relationship between political stability, political institutions, and economic growth. It suggests that political institutions play an important role in forming economic policy and that, even within the group of democratic institutions, different political institutions lead to different policy outcomes. This research can be combined with the two models discussed in this paper to arrive at an interesting direction for further research. The research-guiding question is, Which political institutions are suitable to prevent institutional sclerosis in politically stable countries? Based on the considerations in this paper, suitable political institutions limit the power of distributional coalitions. In addition, they reduce the probability to become caught in the low-growth equilibrium and provide instruments to escape this situation quickly and at low costs should the country become trapped.

I would like to thank Dennis Coates, Jac Heckelman, Lars Ponterlitschek. Dorothea Reimer, Stefan Schafer, Lars Siemers, Lars Tanzmann, Benjamin Weigert, and Bonnie Wilson for helpful comments on an earlier version of this paper.

References

Abramovitz, Moses. 1983. Notes on international differences in productivity rates. In The political economy of growth, edited by Dennis C. Mueller. New Haven, CT: Yale University Press, pp. 79-89.

Ansolabehere, Stephen, and James M. Snyder. 2000. Valence politics and equilibrium in spatial election models. Public Choice 103:327-36.

Armstrong, Harvey, and Robert Read. 1995. Western European micro-states and EU autonomous regions: The advantage of size and sovereignty. Worm Development 23:1229-45.

Bischoff, Ivo. 2003. Determinants of the increase in the number of interest groups in western democracies: Theoretical considerations and evidence for 21 OECD countries. Public Choice 114:197-218.

Breuss, Fritz, and Markus Eller. 2003. Efficiency and federalism in the European Union. Vienna Research Institute for European Affairs Working Paper No. 50.

Butkiewicz, James L., and Halit Yanikkaya. 2006. Institutional quality and economic growth: Maintainance of the rule of law or democratic institutions, or both? Economic Modelling 23:648-61.

Choi, Kwang. 1983. A statistical test of Olson's model. In The political economy of growth, edited by Dennis C. Mueller. New Haven, CT: Yale University Press, pp. 57-78.

Clark, Jeremy. 1998. Fairness in public good provision: An investigation of preferences for equality and proportionality. Canadian Journal of Economics 31:108-29.

Coates, Dennis, Jac C. Heckelman, and Bonnie Wilson. 2006. Determinants of interest group formation. Paper presented at the 2006 Southern Economic Association conference, Charleston, SC. de Haan, Jakob, Susanna Lundstroem, and Jan-Egbert Sturm. 2006. Market-oriented institutions and policies and economic growth: A critical survey. Journal of Economic Surveys 20:157-91.

Denzau, Arthur T., and Douglas C. North. 1994. Shared mental models: Ideologies and institutions. Kyklos 47:3-31.

Easterly, William, and Aart Kraay. 2000. Small states, small problems? Income, growth, and volatility in small states. Worm Development 28:2013-27.

Edmonds, Bruce. 1999. Modelling bounded rationality in agent-based simulations using the evolution of mental models. In Computational techniques for modelling learning in economics, edited by Thomas Brenner. Boston: Kluwer Academic Publishers, pp. 305-32.

Enelow, James M., and Kelvin J. Hinich. 1984. The spatial theory of voting. Cambridge: Cambridge University Press.

Fechtenhauer, Detlef, Alexandra Haferkamp, Franz N. Gresser, and Fabian Christandl. 2006. Efficiency versus justice--On laymen's evaluations of economic policies. Paper presented at the 2006 IAREP-SABE joint meeting, Paris.

Feld, Lars P., and Stefan Voigt. 2006. Making judges independent--Some proposals regarding the judiciary. In Democratic constitutional design and public policy--Analysis and evidence, edited by Roger D. Congleton and Brigitta Swedenborg. Cambridge, MA: MIT Press, pp. 251-88.

Feng, Yi. 1997. Democracy, political stability and economic growth. British Journal of Political Science 27:391-418.

Fuhrmann, Horst. 1996. Uberall ist Mittelalter. Munich: Verlag C.H. Beck.

Gray, Virginia, and David Lowery. 1988. Interest group politics and economic growth in the U.S. States. American Political Science Review 82:109-31.

Groseclose, Tim. 2001. A model of candidate location when one candidate has a valence advantage. American Journal of Political Science 45:862-86.

Haferkamp, Alexandra, Detlef Fechtenhauer, Franz N. Gresser, and Fabian Christandl. 2006. The justice of tax systems What if economic laypersons were themselves ministers of finance. Paper presented at the 2006 IAREP-SABE joint meeting, Paris.

Heckelman, Jac C. 2007. Explaining the rain: The Rise and Decline of Nations after 25 years. Southern Economic Journal 74:18-33.

Heckelman, Jac C., and Dennis Coates. 2003. On the shoulders of a giant. In Collective choice: Essays in honor of Mancur Olson, edited by Jac C. Heckelman and Dennis Coates. Berlin: Springer Verlag, pp. 1-13.

Inman, Robert P., and Daniel L. Rubinfeld. 1997. The political economy of federalism. In Perspectives on public choice, edited by Dennis C. Mueller. Cambridge: Cambridge University Press, pp. 73-105.

Keefer, Philip. 2004. What does political economy tell us about economic development--And vice versa? Annual Review of Political Science 7:247-72.

Knack, Stephen. 2003. Groups, growth and trust: Cross-country evidence on the Olson and Putnam hypotheses. Public Choice 117:341-55.

Knack, Stephen, and Philip Keefer. 1995. Institutions and economic performance: Cross-country tests using alternative institutional measures. Economics and Politics 7:207-27.

Lampert, Heinz. 1997. Die Wirtschafts- und Sozialordnung der Bundesrepuhlik Deutschland. 13th edition. Munich: Olzog Verlag.

Lebas, Adrienne. 2006. Polarization as craft--Party formation and state violence in Zimbabwe. Comparative Politics 38:419-38.

Lindbeck, Assar. 1997. The Swedish experiment. Journal of Economic Literature 35:1273-319.

Maddison, Angus. 1988. Ultimate and proximate growth causality: A critique of Mancur Olson on the Rise and Decline of Nations. Scandinavian Economic History Review 36:25-9.

Maitland, Ian. 1985. Interest groups and economic growth rates. Journal of Politics 47:44-58.

McCallum, John, and Andre Blais. 1987. Government, special interest groups, and economic growth. Public Choice 54:3-18.

Mueller. Dennis C. 2003. Public choice III. Cambridge: Cambridge University Press.

Murrell, Peter. 1984. An examination of the factors affecting the formation of interest groups in OECD countries. Public Choice 43:151-71.

Nardinelli. Clark, Myles S. Wallace, and John T. Warner. 1987. Explaining differences in state growth. Catching up versus Olson. Public Choice 52:201-13.

Norpoth. Helmut. 1996. Presidents and the prospective voter. Journal of Politics 58:776-92.

North, Douglass C. 1990. Institutions, institutional change and economic performance. Cambridge: Cambridge University Press.

Oates, Wallace E. 1999. An essay on fiscal federalism. Journal of Economic Literature 37:1120-49.

Olson, Mancur. 1982. The rise and decline of nations: Economic growth, stagflation and social rigidities. New Haven, CT: Yale University Press.

Olson, Mancur. 1996. Big bills left on the sidewalk: Why some nations are rich and others are poor. Journal of Economic Perspectives 10:3-24.

Paldam, Martin. 2004. Are vote and popularity functions economically correct? In The encyclopedia of public choice, vol. 1. edited by Charles K. Rowley and Friedrich Schneider. Dordrecht: K1uwer Academic Publishers, pp. 49-59.

Paque Karl-Heinz. 1996. Why the 1950s and not the 1920s? Olsonian and non-Olsonian interpretations of two decades of German economic history. In Economic growth in Europe since 1945, edited by Nicholas Crafts and Gianni Toniolo. Cambridge: Cambridge University Press, pp. 95-106.

Persson, Torsten, and Guido Tabellini. 2006. Democracy and development: The devil in the details. American Economic Review Papers and Proceedings 96:319 24.

Posner, Richard A. 1997. Equality, wealth, and political stability. Journal of Law, Economics, and Organization 13:344-65.

Przeworski, Adam, and Fernando Limongi. 1993. Political regimes and economic growth. Journal of Economic Perspectives 7:51-69.

Riechmann, Thomas. 2001. Learning in economics: Analysis and application of genetic algorithms. Heidelberg: Physika Verlag.

Rodrik, Dani. 1996. Understanding economic policy reform. Journal of Economic Literature 34: 9-41.

Rosser, J. Barkley. 1999. On the complexities of complex economic dynamics. Journal of Economic Perspectives 13:169-92.

Sanhueza, Ricardo. 1998. The hazard rate of political regimes. Public Choice 98:337-67.

Strumpf, Koleman S. 2002. Does decentralization increase policy innovation? Journal of Public Economic Theory 4:207-41.

Sunde, Uwe. 2006. Wirtschaftliche Entwicklung und Demokratie--Ist Demokratie ein Wohlstandsmotor oder ein Wohlstandsprodukt. Perspektiven der Wirtschaftspolitik 7:471-500.

Tang, Eddie W. Y., and R. Alan Hedley. 1998. Distributional coalitions, state strength, and economic growth: Towards a comprehensive theory of economic development. Public Choice 96:295-323.

Tetlock, Philip E. 1999. Theory-driven reasoning about plausible pasts and probable futures in world politics: Are we prisoners of our preconceptions? American Journal of Political Science 43:335-66.

Unger, Brigitte, and Frans van Waarden. 1999. Interest associations and economic growth: A critique of Mancur Olson's Rise and Decline of Nations. Review of International Political Economy 6:425-67.

von Hayek, Friedrich A. 1945. The use of knowledge in society. American Economic Review 35:519-30.

von Hayek, Friedrich A. 1972. Die Theorie komplexer Phanomene. Tubingen: Mohr Siebeck.

Weingast, Barry R. 1995. The economic role of political institutions: Market-preserving federalism and economic development. Journal of Law, Economics, and Organization 11:1-31.

Whaples, Robert. 2003. If I had a hammer: Mancur Olson as an economic historian. In Collective choice: Essays in honor of Mancur Olson, edited by Jac C. Heckelman and Dennis Coates. Berlin: Springer-Verlag, pp. 143-61.

Wohlgemuth, Michael. 2002. Evolutionary approaches to politics. Kyklos 55:223-46.

(1) Offering growth-promoting platforms lies within the long-term interest of political parties if the number of elections a party can be expected to win when it offers these policies and is assigned the corresponding high valence is higher than the number of won elections for any other policy platform, including all popular policy platforms. In this case, a party offering growth-promoting platforms can still be defeated by a party offering popular policy platforms in the short run. In the long run, however, a party that has achieved the superior valence associated with good macroeconomic outcomes maximizes its expected votes by continuing to offer growth-promoting policy platforms. Essentially, this means that the majority of voters prefer growth-promoting policies once these are discovered. The previous illustrations implicitly assume that this is the case.

(2) It is important to note two things. First, the same events (e.g., World War II) may affect different countries in different ways and to different extents. Second, the course of argumentation put forth here does not apply to small shifts in economic structure. Small shifts do not devaluate the knowledge about the formerly growth-promoting policies but merely require them to be adjusted. These are the types of adjustments all democratic governments have to cope with in their everyday work.

(3) A similar course of argumentation may apply to small countries and explain why these tend to grow faster than large countries (Armstrong and Read 1995; Easterly and Kraay 2000). Accordingly, voters of small countries tend to know much more about their neighbors and their policy and performance than large countries. Thus, they are able to learn from a larger number of real experiments than voters in large countries.

Ivo Bischoff, Department of Economics, Philipps-University Marburg, Am Plan 2, 35037 Marburg, Germany; E-mail [e-mail].

Table 1. Olson's Chain of Cause and
Effect and the Ways It Has Been Tested

                                        Classification scheme
Hypotheses making up the chain           for empirical tests

1) The number of distributional        I
coalitions within a country         (mixed
increases as the country           evidence)
remains politically stable.

2) If the number of                                              V
distributional coalitions             II                     (largely
is larger, their influence                                  supportive)
increases.                                         IV
                                                 (weak
3) Distributional coalitions                    support)
follow redistributive interests
rather than efficiency-
enhancing ones. If their              III
political influence increases,
economic growth rates will fall.