Formal institutions

Tsebelis (Reference Tsebelis2002) proposes a simple means to compare empirically observable political institutions, which are enormously heterogeneous and usually characterized by rather complex sets of formal rules. This complexity is reduced to one criterion, namely the number of veto players involved in the political decision-making process – which he considers to be the most important parameter to explain political stability. Stability is understood as the constancy in time of both the political aims that the decision-makers define and the measures that are taken in their pursuit. It is important to note that this definition does concern only the set of aims and policies, but not the persons in office. If the political personnel changes, the material content of policies may nevertheless remain unchanged. A veto player is any individual, collective or organization (institution in the political science terminology) whose opposition to a proposed change of the political status quo is sufficient to have the status quo left unchanged. This does imply that the role of a veto player is not necessarily designated through a formal institution, e.g. a constitution laying out the formal process of collective decision-making, but she can also grow into this role informally, for example if she is an influential consultant to the individual or committee formally endowed with the competencies to decide. In this subsection, however, the discussion is restricted to veto players whose influence follows from formal institutions.

On a general level of analysis, the veto player approach allows us to compare a number of ideal types of institutional arrangements for collective decision-making. For example, one can expect presidential democracies to display higher political stability than parliamentary democracies. The reason is that, in the former, it is quite likely that the head of the executive branch and the majority of the legislative branch represent different political parties, while, in the latter, members of government do usually rely upon a majority in parliament provided by their own parties. A more detailed analysis shows then that a parliamentary system where members of parliament are elected through a first past the post majority vote are more likely to generate unambiguous one party majorities, while systems of proportional representation are more likely to necessitate coalition governments in order to organize a majority. The latter increases the number of veto players, which leads us to expect that on average, proportional representation will imply less frequent and less significant political innovations.

The introduction of direct-democratic rights may also play a role. Prima facie, the opportunity to hold a facultative referendum and the duty to hold an obligatory referendum introduce an additional formal veto player, the median voter, which implies less favourable conditions for public entrepreneurship. The instrument of an initiative does, on the other hand, enable the voters to force new items on to the agenda of collective decision-making, even against the opposition of elected representatives, which clearly can be expected to facilitate political innovation.

In distinguishing the innovative propensities of a federalized and a centralized structure of government, veto player theory is also useful. An established argument here is that decentralized jurisdictions can be used as political laboratories in which experiments can be conducted with so far untested sets of policies whose results are uncertain ex ante (see Oates, Reference Oates1999). In addition to this laboratory argument, it can be stated that a federal structure of government does also allow an introduction of additional checks and balances on elected representatives, without having to introduce an additional veto player. Being able to migrate at low cost between relatively small political entities clearly is an instrument for citizens to control their governments. But even without migration, the access to information about the costs and effects of policies conducted elsewhere offers an opportunity for voters to monitor their governments more efficiently (Besley and Case, Reference Besley and Case1995). And beyond this argument in the tradition of an economics of information, the problem can also be analysed from a more evolutionary perspective: the generation of knowledge in a setting of inevitable uncertainty of individuals regarding the concurrence of their fallible hypotheses with the unknown ‘true’ model of the causal relationships that are responsible for the impact of political measures. In such a setting, the decentralized political laboratories can serve as a framework for tests of political hypotheses and help to broaden the base of policy-related knowledge, which may ultimately lead to more rational policies (Heine, Reference Heine2003; Kerber, Reference Kerber and Kurt2004; Schnellenbach Reference Schnellenbach2004a, Reference Schnellenbach2004b).

Summing up, we can therefore expect that formal institutions do exert an influence on the frequency and significance of political innovations – provided that public entrepreneurs exist who indeed have an incentive and the necessary skills to advocate and to promote such innovations. This is still a crucial assumption. If public entrepreneurs do not appear, then political innovations do not enter the political arena in the first place, and the influence of formal institutions becomes irrelevant. As a rule, however, and quite in line with intuition, the introduction of novelty becomes more difficult with the number of veto players, i.e. with the increased presence of formal checks and balances. Beyond the formal institutions, we can also expect informal political institutions to play a role.

Informal institutions

Approaches in the neo-classical tradition of Public Choice generally care very little about informal political institutions. Political results follow from well-defined individual preference orderings that are aggregated according to the rules dictated by an implicitly or explicitly assumed set of formal institutions. Voter behaviour is analysed along similar lines as that of committee members or lobby groups, namely as a maximization problem involving a strictly self-interested utility function and a set of relevant restrictions. Even though one of the grand themes of Public Choice has been the indeterminacy and instability of collective decisions due to problems of aggregation (Schnellenbach, Reference Schnellenbach2005a), individual decisions are fully determined by the utility function and the restrictions under which the individual decides. One of the few exceptions are models of probabilistic voting, where somewhat paradoxically the impossibility to anticipate individual voting behaviour with certainty leads to stable, predictable collective results (Coughlin, Reference Coughlin1992). But even here, clearly defined, self-interested individual preferences define the point of departure.

This stubborn adherence to the assumption of a strictly self-interested individual decision-making calculus is somewhat surprising, given the fact that empirical evidence suggests that voters’ decisions are made in a sociotrophic rather than a self-interested fashion (Mueller, Reference Mueller2003: 460). Voters tend to take general norms of fairness into consideration, and, on average, they even appear to act morally in the sense that they also take into consideration what kind of an effect a proposed political measure will have on the welfare of other individuals. This empirical result does not even necessarily have to appear as an anomaly from the neo-classical perspective. Under the condition of rational ignorance, even a small warm glow as a psychological reward for behaving morally is sufficient to vote along general norms and not according to one's strict economic self-interest (Kirchgässner, Reference Kirchgässner1992).

This reasoning implies that the assumption of informal institutions – understood as not formally codified, but informally communicated political preferences and hypotheses – playing a role in the political sphere is not even inconsistent with a rigorous concept of individual rationality. Following Denzau and North (Reference Denzau and North1994), one could denote such institutions on a more general level also as shared mental models, which have at least two components. On the one hand, they are normative notions shared by a large number of individuals, including for example normative statements on the desired form and extent of government intervention into the market process, and on the desired rules that institutionally define the markets on which individuals actually co-ordinate their actions. More generally phrased, these are shared demand schedules for the provision of publicly provided goods and schemes for their financing. They do not follow from individual constrained maximization, but from societal communication processes. On the other hand, informal institutions also comprehend a positive component: shared hypotheses that enable individuals to understand and interpret political problems, to form expectations regarding which political aims are attainable, and how they are attainable. Informal political institutions therefore constitute a communicated pattern of interpreting and evaluating political reality (Meier and Haudenschild, Reference Meier and Christof1991; Meier and Slembeck, Reference Meier and Tilman1997).

The citizen at the ballot is thus not ‘rational in the Cartesian sense, whereby [he] is able to learn all things and process all information’ (Potts, Reference Potts and John2003: 196). Even if biological evolution had equipped her with the necessary prerequisites to act in this manner, she would not, under this particular set of incentives, have rationally desired to do so. With the utility of holding a scientifically accurate set of positive hypotheses being practically zero, it is reasonable to assume that individuals simply accept the informal institutions communicated to them in their milieu, i.e. to exhibit herd behaviour. In turn, the stability of informal political institutions in time follows from the fact that, once an individual has settled with an informal political institution, she has no individual incentive to invest further costs into learning about alternatives or even into investigating their relative accuracy on her own. A dichotomy between rational and norm-governed actions (Dosi et al., Reference Dosi, Luigi, Andrea and Marco1999; Hodgson, Reference Hodgson1997) is not relevant here; rather, it is entirely rational to decide according to informal political institutions that are communicated in the public discourse.

This is not to imply that all individuals hold the same preferences on public policy and share the same patterns of interpreting political reality. The diffusion of informal political institutions can be plausibly modelled as a frequency dependent process during which the probability that an individual subscribes to a certain informal institution depends positively on the number of other individuals with whom she communicates who already hold this position. An equilibrium will then under plausible conditions be characterized by two or more groups, each consisting of a stable fraction of the population that can be distinguished by the informal political institutions they adhere to. The process therefore bears some similarity to the diffusion of rumours (Banerjee, Reference Banerjee1993) or of informal rules outside of the political sphere (Witt, Reference Witt1989).

From the perspective of every individual, the combination of normative political statements and positive conjectures on issues relevant for policy that she holds, constitutes an informal political institution and guides her actions in the political sphere. Generally, the concept of an institution also encompasses that an enforcement mechanism exists to ensure that individuals do indeed adhere to this institution. The already mentioned effect that no private benefits are to be gained by deviating from a once-accepted informal political institution is one such mechanism. Additionally, one can think of think of peer pressure in social groups, where deviating from accepted patterns of political argumentation is punished through informal sanctions such as exclusion from the group. On the other hand, in a representative democracy informal political institutions also need to be enforced vis-à-vis those representatives who have the formal competencies for decision-making on public goods. An obvious mechanism is the opportunity of a majority of voters to punish representatives who deviate from their informal political institutions at the ballot, but there may also be more subtle mechanisms at work. For example, the threat of exit – taxpayers who threaten to relocate their tax bases to other jurisdictions – may be such a mechanism.

To sum up, informal political institutions as defined here consist of (i) shared positive hypotheses and (ii) shared political preferences that are (iii) self-stabilizing through communication among citizens and (iv) enforced, although often imperfectly, vis-à-vis representatives through both voice and exit, e.g. of tax bases. The degree of enforcement will depend, inter alia, on the interaction with formal institutions that influence the degree of political competition and limit the scope of individual decision-making in the public sector through the provision of formal checks and balances. With its emphasis on informal political institutions, the argument presented here is compatible with a tradition of institutionalist thinking in political theory (March and Olsen, Reference March and Olsen1989). This tradition views politics not as a sequence of individual acts of utility maximization. Rather, institutions that are relatively stable in time and difficult to be manipulated consciously guide the individual and make her ‘familiar with the rules of appropriate behaviour and with the moral and intellectual virtues of the polity’ (March and Olsen, Reference March and Olsen1989: 161) and her political preferences are not given, but evolve through social interaction. Furthermore, these institutions do not need to correspond perfectly with reality: There remains a strict dualism between objective causality and the subjective perception of it (Dopfer, Reference Dopfer and Dopfer2001).

Stylized facts on political innovation

A quintessential insight to be gained from case studies on political innovation is the fact that significant changes of policies, beyond piecemeal accommodations to changed circumstances, almost certainly appear only when the status quo is considered to be definitely unsustainable by the decision-makers themselves and also by a majority of the citizens they represent. Political reforms are rarely forward-looking over a more distant time horizon, and seldom are political innovations implemented as early, precautionary answers to foreseeable difficulties that would occur in the longer run under the status quo policies, or as provident appropriations of attainable welfare gains.

A prominent example is the problem of ground water basins in California analysed by Elinor Ostrom, in her dissertation, and later in Ostrom (Reference Ostrom1990). Ostrom shows that the externality problem of over-exploiting these basins was successfully solved by decentralized water companies themselves, without needing a government agency to enforce an efficient solution. She denotes this as public entrepreneurship, but she does also point out that this solution was only found after a judge had issued a credible threat to enforce a solution to the externality problem that would have been unequivocally disadvantageous to all parties involved. Clearly, both the fact that the status quo has become untenable and the looming of a very unfavourable judicial solution have triggered a process of political innovation. But such a rather passive reaction to a devaluation of the status quo seems to have little in common with the bold, intrinsically motivated pursuit of novelty that the Schumpeterian approach, as well as Potts’ argument, suggests.

In a related argument and taking empirical evidence into account, Kingdon (Reference Kingdon1995) suggests that the timing of political innovations depends crucially on the emergence of windows of opportunity. Public entrepreneurs are characterized as individuals who spend most of their time and resources preparing policy proposals that they can push onto the agenda when windows of opportunity open – but they can do little to influence the emergence of such windows. Michael Mintrom similarly takes the restrictions into consideration that limit the policy entrepreneur's scope for inducing change. He does, however, partly also share the Schumpeterian perspective in his emphasis on personal skills as a prerequisite for public entrepreneurship (Mintrom, Reference Mintrom2000: 113–157 and 268–275). This leads Mintrom to take an intermediate position. He is convinced that public entrepreneurs are not restricted to passively await the emergence of political opportunities. He does, however, also note that introducing novelty requires the investment of a lot of resources and is a tedious enterprise. Compared with the original considerations of Schumpeter, generating public and political support for novel policies is seen here as a much more incremental and also risky task.

Schneider et al. (Reference Schneider, Paul and Michael1995) are also concerned with the responsive nature of public entrepreneurship. They observe that political entrepreneurs are equipped with an alertness which allows them to recognize latent demands for political novelty, and that they are also equipped with skills which allow them to transform the latent into an actual demand. While they acknowledge that political innovators may attempt to exploit the indeterminacy of collective decision-making and to strategically frame collective decisions according to their preferences, Schneider et al. find a regular catalyst of political innovation in changes of political or technological constraints, such as a devaluation of the status quo or changed tastes of the electorate. The numerous case studies presented by Schneider et al. also show that the political innovations they discuss are often not directed at introducing major political changes, but rather at pursuing given political ends in only a slightly different manner.

The responsive nature of public entrepreneurship does also regularly manifest itself in the observation of delayed reforms. There are a number of approaches concerned with this phenomenon, i.e. with the fact that political innovations are implemented not at the time when rational and informed observers expect them to be welfare enhancing, but substantially later. This time pattern has been found to be an empirical regularity in the economics of reform processes (Drazen, Reference Drazen1996), and the explanations offered for this observation are mostly using a game-theoretic framework. Drazen, for example, models the problem as a war-of-attrition game between interest groups. Also, it is often and in concordance with our discussion above assumed that the value of the status quo depreciates in time, so that at some point a sufficient demand for political innovations arises (Alesina and Drazen, Reference Alesina and Alan1991). Rodrik (Reference Rodrik1996) offers a survey of approaches that attempt to explain delayed reforms by looking both at institutions and at bargaining problems between interest groups.

These contributions hint at an important point that is neglected in Schumpeterian approaches to public entrepreneurship: the demand side. A look at prominent figures that are sometimes hailed as prototypes of political entrepreneurs is also enlightening in this respect. Margaret Thatcher, for example, inherited economic policies from her predecessor that have been widely regarded as unsustainable, which virtually forced her to experiment with a different set of policies. And Matsusaka (Reference Matsusaka2004: 100–110) has a closer look at the political career of Ronald Reagan. He finds that in his first years as governor of California, Reagan had been a rather traditional Republican interested in a strong government generously funded with ample tax revenue. Only after observing a strong and growing popular movement for constitutional limits on taxation in California did he sense a change of sentiment among his voters and transformed himself into an advocate of small government and lean bureaucracy. This is another recurring characteristic of political innovation processes: a change on the demand side of the political process precedes a change of the positions taken by elected representatives (see also the numerous cases presented in Kamarck, Reference Kamarck2004).

Political entrepreneurs or political arbitrageurs?

Schumpeter (Reference Schumpeter1942) asserts that the individual level of knowledge on political issues is not to be assessed too amply, and that the conjectures that lead individuals to favour one political measure over the other often have nothing to do with an accurate understanding of the causal relationships that a well-educated specialist would find relevant. His argument does, however, differ from ours in his claim that individual citizens can easily be manipulated by professional politicians. According to Schumpeter, citizens do not tend to preserve and defend political positions once they have accepted them. Recurring to the psychology of the masses, which was influential at the time, Schumpeter writes: ‘Newspaper readers, radio audiences, members of a party even if not physically gathered together are terribly easy to work up into a psychological crowd and into a state of frenzy in which attempt at rational argument only spurs the animal spirits’ (p. 257). He goes on to argue that the political reasoning of the ordinary citizen is ‘associative and affective’ (p. 262) and that professional politicians ‘are able to fashion and, within very wide limits, even to create the will of the people. What we are confronted with in the analysis of political processes is largely not a genuine, but a manufactured will’ (p. 263).

Given the citations above, Schumpeter does not appear to acknowledge the existence of informal political institutions that are stable at least in the short run. On the contrary, the perception of political problems and the assessment of alternative policies are attributed a high degree of volatility. This idea, that it is one of the most important exercises for a politician to influence public opinion, can still be found in some modern contributions to political economics. For example, Witt (Reference Witt2003) argues that setting the agenda in public discourse, convincing colleagues and citizens of one's own preferred policies are the pivotal activities of a professional politician. Wohlgemuth (Reference Wohlgemuth2002) makes a similar point.

Schumpeter, drawing normative conclusions from his approach to political economy, argues not in favour of formal institutional checks and balances to control a potentially erratic political process, but instead argues that one should hope for benevolent, rational public entrepreneurs to make responsible use of their office. This conclusion prima facie appears to be paradox. On the one hand, Schumpeter criticizes the classical theory of democracy, which expects democratic institutions to facilitate an implementation of a volonté générale, for its lack of realism. He does so both on the ground that no such volonté générale exists, and because he expects representatives to be imperfect agents of their principals. But then, he dodges the question of institutional constraints and instead hopes that ‘the human material of politics . . . should be of sufficiently high quality’ (Schumpeter, Reference Schumpeter1942: 290), which in his opinion presupposes the ‘existence of a social stratum, itself a product of a severely selective process, that takes to politics as a matter of course’ whose members are endowed with ‘traditions that embody experience, with a professional code and with a common fund of views’ (p. 291). These statements do, to some extent, dissolve the paradox. Schumpeter appears to be convinced that breeding and selecting a class of professional politicians that is committed to moral and ethical standards – that constrains itself – can solve the problems posed by instable democratic processes. His proposed solution is that ‘electorates and parliaments must be on an intellectual and moral level high enough to be proof against the offering of the crook and crank’ (p. 294).

Implicit in this argument is the assumption that democratic selection mechanisms are efficient enough to install individuals in political offices who indeed have a comparative advantage in the skills that are assumed to be essential for professional politicians, and who also are sufficiently moral characters. If this can be achieved, then control is largely a matter of self-control within the political elite, and formal checks and balances do not need to be considered. Even when Schumpeter demands that ‘the effective range of political decision should not be extended too far’ (p. 291), he apparently does not advocate formal limitations, for example the exclusion of certain unalienable rights from political deliberation. Rather, he hopes that representatives will be prudent enough to ‘accept the specialists’ advice whatever they may think themselves’ (p. 292) on issues where their own expertise does not suffice to make well-informed decisions. But if all this were indeed the case, i.e. if at least some democracies tended to bring forth a trustworthy, self-constrained but also entrepreneurial political elite, then the universal phenomenon of delayed reforms could hardly be explained: Why should representatives, who have the skills to implement welfare-enhancing (with appropriate schemes of redistribution also pareto-improving) political innovations, abstain from doing so?

One possible answer to this question is the reference to the concept of informal political institutions introduced here. It has already been argued above that the possibility of informal sanctions as a response to deviating individual political positions within a group or a milieu can be expected to stabilize a given distribution of informal political institutions within a population. But there are other mechanisms beyond punishment of dissenters that also contribute to such a result. Individuals tend to unconsciously preserve their given set of (positive) conjectures by looking for feasible solutions to political problems through the filter of these conjectures (Dopfer, Reference Dopfer1991). If they find a policy that yields acceptable results, then this result confirms the set of conjectures that has brought forth this policy. A necessary condition for an incentive to critically examine one's conjectures is that the political status quo becomes sufficiently unattractive and the conjectures upon which it rests become discredited. But even this is not necessarily a sufficient condition for individual learning processes to commence. Loury (Reference Loury1994) argues that processes of communication within a collective of people are strongly characterized by incentives to ignore information and argumentation that casts doubt on the common sense that is currently prevailing within the group. There is a strong tendency towards a not formally enforced self-censorship in public discourse that, as long as the results associated with the status quo are just acceptable, makes it difficult for dissenting positions to find broad attention.

It is only when the self-confirming communication within social networks ceases due to widespread disenchantment with the status quo that public entrepreneurship is most likely to be successful. At this point, a large number of individuals suffers from a substantial uncertainty regarding the potential of the given informal political institutions to guide policy to acceptable results, and demands either reassuring information or novel propositions. With the number of disenchanted individuals being sufficiently large, reassuring information is not supplied in sufficient amounts to stabilize the informal political institutions and novelty can be successfully introduced into the political process (Schnellenbach, Reference Schnellenbach2005b). Typically, and in accordance with the stylized facts presented above, political innovations will therefore occur in times of crisis and not be implemented with ample foresight by bold public entrepreneurs.

Seen from this perspective, the activities of a public entrepreneur have little in common with the Schumpeterian proposition that public entrepreneurship is merely private entrepreneurship transferred to the political stage. On the contrary, the activities of political representatives appear to justify the label ‘public arbitrageur’ much more than the label ‘public entrepreneur’. The stylized facts discussed above suggest that the activities of political representatives are directed at understanding which informal political institutions restrict them and at setting the agenda according to these restrictions. If the representative wants to actively alter the informal institutions themselves, she will find herself restricted to few windows of opportunity – points in time where the status quo policies and the informal political institutions associated with them are discredited. As long as democratic control or other means of punishing representatives for executing policies beyond the scope of informal political institutions are at work, it is in the self-interest of office-seeking politicians not to deviate from them, or at least not to deviate from them too significantly.

The argument made here implies that with checks and balances in place, the point in time where non-incremental political innovations take place will, as a rule, not be determined by a Schumpeterian public entrepreneur. Rather, the timing of political innovations is determined by the issues. If informal institutions deteriorate in times of crisis, because given perceptions of political problems and their solutions prove to be unfit, then a political entrepreneur may use this window of opportunity to supply novel solutions. But she then acts as a Kirznerian rather than a Schumpeterian entrepreneur, and is most importantly characterized by an altertness for a latent demand for novel solutions (see Kirzner, Reference Kirzner2000: chapter 13 for a comparison of the two perspectives of entrepreneurship in general).

Obviously, this is not to say that Schumpeterian public entrepreneurship is entirely impossible. In a world with heterogeneous individuals it is of course possible, although not likely, that once in a while an exceptionally charismatic person or a very talented demagogue enters the stage and is able to convince a sufficient number of individuals of novel political aims and measures, or even persuades them that given informal political institutions, although not discredited so far, need to be given up in favour of new ones. But these are scarce exceptions that are, as will be argued in the next section, associated with substantial difficulties from a normative perspective. As a rule, one will find public arbitrageurs rather than public entrepreneurs as long as the status quo policy is not discredited.