Emergence and Persistence of Inefficient States

Inhalt

1. Introduction
1.1. Motivation
1.2. Some Historical Observations
1.3. Benefits of a Political Economy Approach
1.4. Structure of the Thesis

2. The Political Economy of Inefficient Institutions
2.1. Setting the stage: State Capacity, Inefficiency and Bureaucracy
2.1.1. Inefficient Institutions
2.1.2. Institutional Change
2.1.3. State Capacity
2.1.4. Bureaucracy
2.2. Inequality, Democratic Transition and the Median Voter Puzzle
2.2.1. Initial Inequality and Political Conflict
2.2.2. Democratic Transition and Inequality
2.2.3. The Median Voter Puzzle
2.3. Related Literature
2.3.1. Why Societies End Up with Inefficient Institutions
2.3.2. Literature on Inefficient States
2.3.3. A new Perspective on the Emergence of Inefficient States
2.4. Empirical Evidence

3. Emergence of Inefficient States under Patronage Politics
3.1. Economic and Political Framework
3.1.1. Tax Evasion, Effort Decision and State Organization
3.1.2. Political System
3.2. Timing of Events
3.3. Equilibria
3.3.1. Constraints and the Government Budget
3.3.2. The Comparative Cases: Permanent Democracy and Nondemocracy
3.3.3. Emerging Democracy: an Inefficient Equilibrium
3.4. Comparative Statics
3.4.1. Characteristics of an Inefficient State
3.4.2. Inequality and Inefficiency
3.4.3. Nondemocracy, Democracy or Emerging Democracy?

4. Discussion and Extension
4.1. Critical Assumptions and the Inefficient State
4.1.1. Agents and Democratic Transition
4.1.2. Public Expenditure
4.1.3. Taxes
4.1.4. Preferences and Inefficiency
4.2. More Comparative Statics
4.2.1. Excessive Bureaucracy
4.2.2. Tax Rates and Inefficiency
4.3. Extensions
4.3.1. The Middle Class and the Inefficient State
4.3.2. Progressive Income Taxation
4.3.3. Endogenizing Inefficiency
4.3.4. Distribution Costs
4.4. Potential and Limits
4.4.1. Explaining Inefficient States
4.4.2. Limits
4.4.3. Contributions to Related Research
4.5. Persistence
4.5.1. Determinants of Persistence
4.5.2. When Inefficient States Persist

5. Concluding Remarks

A. Appendix Proofs

References

Chapter 1. Introduction

Bureaucratic inefficiency, patronage-induced overstaffing, and outright corruption retard economic development and reduce public well-being in developing countries. They prevent governments from effectively carrying out the economic plans to which they devote so much official attention, and deprive citizens of government services to which they are legally entitled. (Geddes 1991, p.371)

Corruption, patronage and inefficient bureaucracies are widespread in many countries and under all forms of government. Whereas such problems are often recognized to be at the core of rent-seeking regimes like dictatorships or oligarchies, they occur in similar ways in developing and even in developed democracies. For instance, there is evidence for bureaucratic inefficiency, corruption and patronage in Brazil (Weyland 1996), but also in Germany (Braendle 2012) and France (Facchini and Melki 2013)

A rich body of research however underlines the importance of good institutions as main determinants of economic development, growth and prosperity. But why and under which conditions do some societies choose inefficient institutions? This is the main question the thesis seeks to answer.

A related second question concerns a possible explanation for the setup of inefficient institu- tions: Why do democracies often pursue relatively elite-biased policies contrary to the predictions of the median-voter-theorem? As Acemoglu (2010, p.1) notes, institutions are often “structured to maximize the rents captured by the elite, the individuals or social groups monopolizing political power". In electoral democracy, the influence of different social groups should be closer related to their population share than in nondemocracy. However, the distribution of political power can by diluted if institutions reshape political influence and redirect government spending towards the top.¹

The thesis tries to bring together the two issues by elaborating a possible explanation for the link between the emergence of inefficient institutions and the prevalence of pro-elite policies. The rationale has first been formulated in a political economy model of inefficient states provided by Acemoglu et al. (2011). Based on the model framework, the main purpose of my analysis is to show under which conditions inefficient states emerge and persist even if alternatives are available at no extra cost.

The general rationale of the baseline model reads as follows. In the initial state, a rich elite is in power, but faces transition to democracy.² Since the elite expect losses in economic and political power after the transition (e.g. through redistributive policies), they try to shape the institutions of the future democracy to their favor. Concretely, the elite consider setting up an inefficient (i.e. unnecessary large) bureaucracy, which is required for taxation, as a result of strategic considerations to keep their political power. With the political support of a large bureaucracy depending on rents provided by the elite, the latter are able to capture democratic politics and thereby reduce redistribution and political power dispersion. In the sense of de Mesquita et al. (2003), bureaucrats form a “winning coalition" with the elite. Since the rich can credibly commit to the provision of rents to bureaucrats, as long as the latter continue to support the party of the elite, the inefficient state persists.

Acemoglu et al. (2011) show that the elite will eventually create an excessive bureaucracy to remain in power. Furthermore, their analysis implies that greater income inequality, intermediate bureaucratic rents and sufficiently forward-looking individuals increase the probability of an inefficient state to arise.

While the model is rebuilt in the third chapter of the thesis, I will first assess the potential of different models explaining inefficient institutions to justify the choice of Acemoglu et al. (2011). Subsequently, the fourth chapter provides an in-depth discussion, comparative statics and several extensions aimed at explaining what causes the emergence of an inefficient state, and under what sort of conditions it arises. In particular, I extend the model by the introduction of an entrepreneurial middle class. The analysis is concluded by an evaluation of the persistence of the model equilibria.

I find that the model is able to explain the emergence of an inefficient state if the elite dispose of a critical size and if relative income inequality as measured by the productivity differential of the poor and the rich is high. However, elite size has an opposite influence on the likelihood of the emergence of an inefficient state. Consequently, an inefficient state only emerges under certain inequality conditions, which provides an answer to the question why certain societies choose inefficient institutions, while others do not.

The elite’s decision to create an inefficient bureaucracy is not an absolute decision motivated by the retention of power, but depends mainly on the relative attractiveness of limited redis- tribution compared to its costs. In the baseline model, simplifying assumptions like a linear tax system, limited liability for tax evaders and equal public good provision work in favor of the emergence of an inefficient state since they systematically decrease the attractiveness of an efficient democracy.

Furthermore, I show that the inefficient state only occurs at intermediate levels of inefficiency, and that higher productivity of the poor up to a certain point leads to higher revenue extraction by the rich, which increases the likelihood of the emergence of an inefficient state. Introducing a middle class into the model, I identify a conditional monotonic relationship between the size and the productivity of a middle class on the one hand and the emergence of an inefficient state on the other. A relatively large and affluent middle class reduces the likelihood of an inefficient state to emerge, while a smaller, relatively productive middle class has an opposite impact. I start my thesis by giving a short theoretical and empirical motivation for the approach.

1.1. Motivation

The approach of the baseline model comprises at least two new perspectives on the emergence and persistence of inefficient institutions. First, the seed of inefficient institutions might already be planted in an early phase of democratization. With the use of political patronage and the creation of an insider/outsider society, a rich elite might capture the political process from the beginning. Knowing that democratization and demand for redistribution require setting up a bureaucracy, the elite exploit this necessity. This is important, since economic and political institutions evolve endogenously in interaction with their economic environment over time, but depend on initial economic conditions (Acemoglu et al. 2004). Yet if the setup of bureaucracy results from an inefficient political agenda and not out of pure technical necessity, this can predetermine the path to an ever-growing bureaucracy.

Second, in the chosen model the inefficiency of the state is not considered to be due to internal dysfunctional processes, slow adaptation or monitoring incapacity, but due to strategic purposes of a small elite, which is fully aware of the explicitly created inefficiencies.³ The rationale implies that all “internal" or procedural explanations of inefficient institutions only capture the outcome of an underlying political process. Consequently, also the persistence of inefficient institutions might rather be the result of a conscious political choice than due to organizational or accidental reasons or changing environments (as suggested by the developmental view of institutions).

However, with a structural focus on patronage politics and bureaucracy and a temporal focus on emerging democracy, the baseline model only displays an isolated mechanism that potentially creates persistent inefficiency in a certain phase of democratic transition. Naturally, there are numerous other possibilities for the elite to act in order to keep political power in democracy. These include lobbying, media control, (direct) vote buying, use of (military) force and co- optation.⁴

Still, the specific focus of the model is justified: It will become evident that the actual size of the required model-bureaucracy to win an election may be too large, and the elite will hence use the discussed possibilities as complementary measures to support their political campaigns. However, the proposed top-down patronage mechanism should be discussed in theoretical isolation as it best captures underlying structures that pave the way to the creation of an inefficient state. Moreover, the mere focus on bureaucracy will provide important insights into inefficient states. A rich body of literature suggests that low redistribution, low institutional quality and excessive bureaucracy go hand in hand.⁵ The lack of state capacity (e.g. the inability to raise taxes and an ineffective bureaucracy) has been shown to be an important determinant of the emergence of inefficient states. Limited state capacity is a good proxy for inefficiency - both in terms of suboptimal output and corrupt bureaucracy.

Autonomous bureaucracies hence seem to be a main determinant of growth and prosperity. As Max Weber noted in Economy and Society, “Capitalism and bureaucracy have found each other and belong intimately together" (Weber 1968, p.1395). Yet ideal-typical Weberian bureaucrats, who are only concerned to fulfill their role in the bureaucratic apparatus, are rare in reality. More common is a partial or complete absence of a coherent bureaucratic apparatus, a clear impediment to development and a source of vast inefficiencies (Evans 1992, p.146). Consequently, the chosen political economy model linking bureaucracy, patronage politics and inefficient institutions seems to be a promising and fruitful approach.

1.2. Some Historical Observations

After having stated the theoretical advantages of the model approach to inefficient states, it remains to provide some empirical and historical motivation. A typical example stated by Acemoglu et al. (2011) is the case of Brazil, where patronage relations in the public sector have ensured the political power of the elite and at the same time led to a limited amount of public good provision. Interestingly, even the poorest social groups have repeatedly supported traditional parties instead of parties opting for higher levels of public good provision (Gay 1990).

Still, the literature shows that other Latin American states also have a doubtable record of excessive and patronage-driven bureaucracy. As Geddes (1991) notes, government agencies in Uruguay during the 1950s and 1960s spent nearly their complete budgets on wages for bureaucrats, such that operating expenses, e.g. for public hospital equipment, could not be covered. Opening a small business in Peru could take up to one year and involve repeated bribery due to 11 different licenses and certificates required (ibid.). This evidence underlines how further inefficiencies such as high entry barriers and reduced government budgets for public goods arise from excessive patronage bureaucracy: Once installed, bureaucracy may be used for political and economic purposes of the elite.

But also in numerous developed countries with long democratic traditions, growing govern- ment expenditure has been accompanied by a disproportional increase of government size and scope for the past 30 years (Persson and Tabellini 1999). Besley and Coate (1998) show that this increase seems to be independent of tax systems or organization of the state. Whereas such observations are not exclusively related to inefficient bureaucracies created for strategic reasons in democratic transition, at least part of the patronage-based reasoning seems to be at work.⁶ For instance, Greece had a system of five administrative levels, a captured bureaucracy apparatus and more than a third of its working population employed by the state before the breakout of the Eurozone crisis in 2009 (Papaconstantinou et al. 2013; Katsimi and Moutos 2010). Also, a so called Eurozone-“core"-countries like France currently employs 90 government officials per 1000 inhabitants, and is often blamed for its large public sector share (Insee 2011).⁷ Even if these excessive bureaucracies might have emerged for different reasons, strategic consider- ations of economically powerful groups surely play a role by now. An approach explaining inefficient states on the basis of strategic political considerations can be motivated by historical considerations.

1.3. Benefits of a Political Economy Approach

To conclude this introduction, I will shortly lay out the benefits of a political economy approach to the problem of inefficient states. Speaking with Persson and Tabellini (2002, p.15), “Political Economy starts with the problem of choice in a society with heterogeneous agents, but with a very different focus than multi-agent welfare economics. The focus is on the process by which it is decided what policy to adopt and, more specifically, on what policy choice will emerge from specific political problems."

A political economy approach thus allows combining the consideration of economic incentives and the economic framework with the political decision process as well as the analysis of their interdependence. Particularly, since political institutions and social conflicts are of primordial importance for economic outcomes in the context of inefficient states, a theory explaining the way institutions are shaped and pointing out what hinders democratic development relying on economic incentives offers the most comprehensive methodical toolkit.

The political economy framework allows for consideration of social and economic systems at the same time. A mere focus on public choice or growth theory would not be sufficient to understand why inefficient states and huge differences in the amount of bureaucratic corruption exist and persist. Most important, these approaches often cannot explain why an inefficient state emerges in some democracies, whereas it does not in others, which is a core advantage of the differentiated political economy approach elaborated in this thesis.

1.4. Structure of the Thesis

The thesis is structured as follows. In the following chapter, I will embed the approach into related strands of research. Central notions like state capacity, redistribution and democratization will be examined concerning their role in the emergence of inefficient states. Furthermore, I will discuss similar models, assess the differences to the model by Acemoglu et al. (2011) and justify its choice for the deeper analysis in the remainder of the thesis.

In the third chapter, I rebuild the model and explain the main insights and comparative statics provided by Acemoglu et al. (2011). An in-depth discussion of model assumptions and outcomes constitutes the first part of the fourth chapter. Subsequently, I present some more comparative statics and provide several extensions before concluding the discussion by an assessment of the potential and the limits of the model in explaining inefficient states. The fifth chapter provides a conclusion and an outlook on possible new research areas.

Chapter 2. The Political Economy of Inefficient Institutions

What are determinants of inefficient institutions? The chapter starts by answering this question with the introduction of basic concepts and terms of the political economy of inefficient institu- tions. Subsequently, I will discuss notions central to the structural foundation of Acemoglu et al.

(2011) to line out and justify the model approach, and to point out related open research questions for the analysis of inefficient states. Equipped with a terminology and a general theoretical framework of inefficient states, I will then discuss different approaches to inefficient states. The section closes with an assessment of the differences and weaknesses of the discussed approaches in comparison to the model by Acemoglu et al. (2011). I conclude the chapter with a short sketch of empirical evidence on the matter of inefficient states.

2.1. Setting the stage: State Capacity, Inefficiency and Bureaucracy

To start with, I focus on different views on institutions and implications from institutional change followed by a discussion of the role of state capacity and bureaucracy in inefficient institutions.

2.1.1. Inefficient Institutions

The importance of institutions for economic development, democracy and growth has been discussed by many scholars and led to the foundation of the new institutional economics (Coase 1998). The often-cited definition from North’s book Institutions, Institutional Change and Economic Performance describes institutions as “rules of the game" and in the broadest sense, the way how society is organized (North 1990). Institutions can reduce uncertainty and transaction costs, while enabling societies to take advantage of economic opportunities.

2.1. Setting the stage: State Capacity, Inefficiency and Bureaucracy

Two characteristics of institutions are of major importance in the context of the emergence of inefficient states. First, institutions provide solutions for commitment and enforcement problems. A government often cannot credibly commit to certain policies - but it can be controlled by regular elections and a legal framework. Commitment to such de jure political institutions determining the distribution of political power plays an important role in democratic transition (Acemoglu et al. 2004; Acemoglu and Robinson 2000). Second, institutions are not necessarily designed efficiently in terms of growth and output, but may be shaped to maximize returns to politically or economically decisive groups. For instance, North observes a “persistent tension between the ownership structure which maximized the rents to the ruler and an efficient system that reduced transaction costs and encouraged economic growth." (North 1981, p.25). Even worse, the problem can be aggravated by the fact that once installed, institutions tend to persist (Coate and Morris 1999).

Consequently, it is clear that institutions can emerge in an inefficient way due to commitment problems. I will now define what is understood by the notions of efficient and inefficient institutions.¹ To start with, a useful formal distinction between economic and political institutions has to be made. The former comprise property rights, contract enforcement and shape economic incentives. The latter refer to forms of government, constraints on politicians and hence regulate the allocation of de jure political power.

Development-promoting efficient institutions are inclusive in the sense that they allow for broad political and economic participation. For instance, Acemoglu and Robinson (2006) list secure property rights, law and order, the provision of (basic) public goods, but also general access to education and opportunity as examples for good economic institutions. Inclusive political institutions enable pluralism, constraints on checks and balances and hence lead to the limitation of social conflicts in a society (Cervellati et al. 2008).

On the contrary, inefficient institutions are extractive and generally characterized by imbalances between particular economic and political interests. Powerful social groups use inefficient economic institutions to extract resources and rents from the rest of the society and prevent or engage in excessive redistribution (Acemoglu and Robinson 2006). Through highly distortionary taxation, they can also engage in factor price manipulation (Acemoglu 2010). Extractive political institutions allow for a concentration of political power in the reigning social group. Lack of constraints and control can lead to patronage, corruption and an inefficient bureaucratic apparatus. As discussed in the introduction, this is the kind of inefficient institutions the model analyzed by the thesis focuses on.

2.1.2. Institutional Change

After having clarified the term of inefficient institutions, I will now turn to the question of institutional change. Naturally, the way (inefficient) institutions can emerge depends on a dynamic view of institutional development. In fact, since most institutions have developed over time and are often path-dependent, it has to be specified how and under which conditions they emerge (North 1990).

Thes days, institutional change is seen as a key determinant of economic and political develop- ment. In particular, democratization has been modeled as an outcome of equilibrium institutional change (Acemoglu and Robinson 2001b). Following this approach, political institutions can lock in de facto political power, which would else stay transitory. During the process of power allocation in transitory phases of democratization, the scope for institutional change is often large. However, it depends on the general view of institutions and their potential for change if this power struggle is assigned to potentially give birth to inefficient institutions. Most in line with the underlying rationale of this thesis is the rent-seeking-institutions approach. Contrary to the efficient institutions view, under which societies automatically choose institutions that maximize total surplus, according to the rent-seeking-institutions approach institutions emerge as a consideration of economic agents’ decisions and are hence not necessarily efficient (Congleton 2004). Moreover, institutions are not necessarily chosen by the whole society, but can also be chosen by specific social groups disposing of political power.² This is exactly how the emergence of inefficient institutions is seen in the model focused on in Chapter 3. It has to be noted that there is also a mere temporal approach to the emergence of inefficient institutions. Since social groups may invest in particular sets of institutions, it might become costly for them to replace or adapt them to a changing economic environment. Consequently, formerly efficient institutions might become inefficient endogenously through delayed reforms and changing economic conditions. Yet for the analysis provided here, the focus rather lies on changes in political than in economic circumstances. Still, parts of this argumentation are used to illustrate how inefficient institutions can persist (see section 4.5).

Starting from these notions of inefficient institutions and institutional change, I will now explicitly discuss state capacity and modern state institutions like bureaucracy on which the Acemoglu et al. (2011) model focuses.

2.1.3. State Capacity

As outlined in the introduction, the model framework is mostly concerned with the efficiency of the organization of the state. A closely related notion in this context is state capacity, which comprises capacities to raise taxes, enforce contracts, regulate markets and the general ability to implement a range of policies (Tilly 1990; Besley and Persson 2009).

Economies with limited state capacity redistribute less, exhibit less growth and are more likely to have corrupt bureaucracies (for instance, see Evans 1995; Besley and Persson 2009; Dincecco and Katz 2012). Consequently, low state capacity seems to be correlated with inefficient institutions (Besley and Persson 2010).

Furthermore, scholars have stressed the interdependence of state capacity and development, and high state capacity is seen as a key to economic development (Knutsen 2013; Acemoglu et al. 2014; Besley et al. 2013). Yet Besley and Persson (2009) show how policy choices can be constrained by past investments in legal and fiscal capacity, and how state capacity evolves over time. They further argue that lack of state capacity is detrimental to economic development as investment in public goods is discouraged, leading to a downward spiral on investments in state capacity. For instance, African and Latin American countries’ economic difficulties have been associated with underinvestment in state capacity (Herbst 2000; Centeno 2003). However, higher state capacity is not necessarily beneficial for society as the value of state control increases, which in turn might induce greater political conflict (Acemoglu 2010).

Now, if state capacity is important, why are persistently low investments in state capacity observed among many countries? Moreover, what determines the level of state capacity? Proximate causes are given by Besley and Persson (2009) and Cardenas and Tuzemen (2011), including the level of wealth, the distribution of political and economic power, the demand for public goods and the economic power of the ruling group. However, the model framework discussed here focuses on underlying processes. In Chapter 4, it will be shown that the model offers a comprehensive explanation for the problems related to state capacity which were discussed in this section.

2.1.4. Bureaucracy

A state’s ability to regulate markets, enforce contracts and provide public goods depends on a coherent bureaucratic apparatus, in which bureaucrats’ self-interests should be aligned with national interests. A government needs bureaucrats to carry out their credible commitment to enforce contracts, which has often been described as a main driver of development (North 1981; Tabellini 2007). Hence, bureaucracy is among the most important institutions as it affects nearly every sphere of human living on earth (Downs 1965). Putting it at the core of a model of inefficient institutions should hence be a valid approach, which the remainder of this section will show.

It does not come as a surprise that bureaucratic quality has an impact on economic growth. For instance, if bureaucratic oversight is weak, neither budgets nor efficiency of bureaucratic work can be controlled. This potentially leads to corruption, which induces deadweight losses and lower growth (see Sarte 1998; Mauro 1996). As mentioned in the introduction, bureaucrats in many economies are however not closely bound by the rational-legal framework that Max Weber developed. Instead, they might consider only interests of certain social groups. Poulantzas (1978) suggested that the bureaucracy better defends the interests of capitalists than they would be capable of on their own.

These technical, outcome-related issues have been widely discussed. It is evident that a certain degree of inefficiency in bureaucracy is inevitable: It would simply be too costly for a society to create perfect monitoring for bureaucracy (Acemoglu and Verdier 1998). Yet this only partially explains why bureaucratic corruption is widespread in many economies, why bureaucracies tend to be oversized and why extensive bureaucratic procedures often impede or delay investment.

The approach to bureaucracy adopted in this thesis emphasizes the strategic underlying processes leading to inefficient bureaucracies.³ Considering the above noted observations, it asks who benefits from an oversized bureaucracy, which directly depends on the government’s budget decisions, and bureaucratic delay, which works like an entry barrier. The rationale is straightforward: If an economically powerful elite anticipates benefits from certain outcomes of an inefficient state organization, they might try to bring about an inefficient bureaucracy.

2.2. Inequality, Democratic Transition and the Median Voter Puzzle

After having outlined the approach to inefficiency and bureaucracy, I will now discuss the structural foundations of the environment in which inefficient institutions may emerge. These include democratic transition and the presence of a certain level of initial inequality and political conflict. The section concludes with some remarks on what I call the “median voter puzzle", i.e. the observation that redistribution in many countries is lower than expected by the median voter theorem.

2.2.1. Initial Inequality and Political Conflict

Initial inequality is an important determinant of policy choice (Robinson 1999). Inequality induces political conflict and has been shown to negatively influence economic growth through various channels (see Galor and Zeira 1993; Alesina and Rodrik 1994; Deininger and Squire 1998). Consequently, as argued by Bénabou (1996), particularly initial inequality determines a state’s political, economic and institutional development. For instance, high initial inequality is prone to reinforce itself through rent seeking activities or a high concentration of political power.

Consequently, in the model approach, initial inequality serves as a basic precondition for the emergence of inefficient states since it induces political conflict. The general rationale here follows Romer (1975) who first formulated the relationship between income inequality among voters and government redistribution. Higher initial inequality should lead to higher redistributive pressure, and hence to more political conflict and more redistribution. While this result has been questioned (see section 2.2.3), the model discussed in Chapter 3 relies on the first part of this mechanism.

2.2.2. Democratic Transition and Inequality

In the structural framework adopted here, democratization happens for exogenous reasons. Equilibrium institutional change is not modeled explicitly to simplify the approach.⁴ Yet several induced assumptions concerning the distribution of political power and resources in the model economy can be derived and tested for their influence on the emergence of an inefficient state (see Chapter 4).

As Shin (1994, p.151) notes, it is difficult to determine general preconditions for the emergence of democracy. Rather than being determined by one single factor, it is more a result of a combination of causes. Furthermore, emergence and consolidation of democracy are linked to different mechanisms. For instance, different paths can be taken from nondemocracy to democracy, some of which gradually lead to full franchise extension (Britain, see North and Weingast 1989), whereas others end up in unconsolidated states (Argentina, most of the 20th century, see Acemoglu and Robinson 2006).

Various approaches to democratization have been discussed. In the model discussed here, the underlying approach follows the Acemoglu and Robinson (2001a) framework, acoording to which democratization happens as a strategic choice of autocratic elites (Rosendorff 2001; Lizzeri and Persico 2004). The choice is motivated by a revolutionary threat to prevent social unrest and a general conflict over redistribution (Weingast 1997; Boix 2003). Democratization hence happens from above responding to a pressure coming from below.

Some direct and some induced implications can be inferred of the assumed unavoidable transition to democracy for the initial conditions in the discussed model of inefficient institutions. As to the direct implications, significant social conflict and some sort of revolutionary threat are basic conditions for democratization to happen in the assumed way. Second, as democratization is supposed to happen in the near future, this threat is not transitory, which implies the inability of the elite to absorb the revolutionary pressure and significant political activity of the poor.⁵ Concerning the induced implications, it can be inferred that repression does not constitute a viable option for the elite to stay in power. Furthermore, the elite need a commitment device, which implies that lower concessions than full democratization are not credible.

The level of initial inequality poses a particular problem. In the Acemoglu and Robinson (2006) framework, democracies should emerge at intermediate values of economic inequality (inverted-U relationship). Democratization is seen as a function of the elite’s fear of the costs of redistribution. However, other approaches based on asset specificity and the struggle of social groups for protection from expropriation find that rather higher income inequality should be monotonically associated with democratization (Ansell and Samuels 2010). Since this problem is linked to the foundation of the model on the median voter theorem and a “redistributionist" rationale, I conclude this section by a closer look at this issue.

2.2.3. The Median Voter Puzzle

Conventional political economy often assumes the median-voter theorem to hold in election outcomes. A lot of models hence incorporate a continuous and monotonic relationship between individual income and the preferred level of redistribution and taxation. Consequently, higher inequality or an extension of the franchise should increase the preferred tax rate of the median voter and result in greater government redistribution.

Yet these predictions seem not to hold in practice. Breyer and Ursprung (1998) even speak of a “paradox of redistribution": In many economies, high inequality and lower-than-expected levels of redistribution are observed. In more drastic words, Ansell and Samuels (2010, p.3) conclude that “democracies do not redistribute more than autocracies and inequality might not be correlated with pressures or redistribution". In fact, the empirical evidence is inconsistent: Some studies provide support for the theorem (Meltzer and Richard 1983; Bénabou 2000; Alesina et al. 2001; Scervini 2012), whereas others find no clear relationship between inequality, franchise extension and redistribution (Clarke 1995; Perotti 1996). Hence, scholars have tried to relax median voter assumptions and to explain outcome-diluting mechanisms. One can basically identify three approaches: First, the median voter theorem has been blamed for its strong underlying assumptions like single-peaked and one-dimensional preferences. Second, technical issues have been brought forward, claiming that the decisive voter is not at the median of the electorate, but related to a different universe. For instance, the voter might not be aware of the level of inequality (information asymmetries), and low turn-up rates of the poor might alter the electorate’s median preference for redistribution.⁶

A more structural third approach deals with the preferences of the median voter and the influence of institutions. For instance, low institutional quality or strategic voting might alter redistribution preferences (Hassler et al. 2003). The implications from the present model framework fit into this line of argumentation. After having rebuilt and extended the model, it will become clear that it provides a different explanation for the failure of the median voter theorem’s predictions. To sum up, it can be noted that the model approach discussed in the subsequent chapter is embedded in a theoretical framework of rent-seeking institutions, where inefficiencies are generated dynamically and consciously during a phase of institutional change caused by political transition. The high relevance of modern state institutions and the importance of state capacity for efficient, prosperous states provide major motivations for the focus of the model rationale on bureaucracy. The assumption of democratic transition as unavoidable sets the stage for the emergence of new political institutions. Lastly, the model relies on the assumptions of the median voter theorem but nevertheless provides a rationale for the inconclusive empirical evidence.

2.3. Related Literature

I now turn to the related literature and to different explanations of inefficient states to further embed the approach adopted in the thesis. Subsequently, I will examine the weaknesses of these approaches in comparison to the Acemoglu et al. (2011) model.

2.3.1. Why Societies End Up with Inefficient Institutions

There are different explanations how a society might end up with inefficient institutions. The rationale within the model discussed here relies on commitment problems and strategic action of a ruling elite. It is partially in line with Acemoglu (2005), who claims that interaction of initial institutions and critical junctures in a historical perspective determine institutional outcomes. More important, the rationale emphasizes the influence of commitment and interaction between social groups in the emergence of institutions. This view is best summarized as follows: Inefficient institutions will emerge and persist [...] when groups that prefer the ineffi- cient (non-growth enhancing) policies that these institutions generate are sufficiently powerful, and when other social arrangements that compensate these powerful groups, while reaching more efficient allocations, cannot be found. (Acemoglu et al. 2006, p.342)

However, it need not be a strategic choice of the elite to enhance the emergence of inefficient institutions. But as discussed in what follows, a lot of approaches to inefficient institutions discussed in the literature are derived from commitment problems of social groups and govern- ments.

First, technical issues like enforcement or monitoring problems have often been hold accountable for inefficiencies - yet their impact can be controlled by ruling social groups (see i.a. Barro 1973). Second, setting up and changing institutions is costly (Alesina and Drazen 1991), and different social groups would need to agree on who bears these costs - which again constitutes a collective action problem. Third, initial economic conditions like natural resource abundance have been associated with the outcome of inefficient institutions, since under these circumstances payoffs from inefficient policies might be higher (Acemoglu 2005; Polterovich et al. 2008). Again, the inefficient outcome is related to strategic behavior of economically powerful groups choosing rent-seeking institutions to their favor.

2.3. Related Literature

The only genuinely different approach to inefficient institutions compared to the model to be discussed relies on time inconsistency. Scholars have argued that institutions become inefficient over time due to their inability to adapt to changing economic circumstances. Path dependency clearly plays a role, but also here, the inefficiency remains linked to relations between social groups: Those who benefit from inefficient institutions - which are often those who could change them - would lose economically and probably politically from institutional change (Coate and Morris 1999). On the other hand, those who suffer might fear repression and face cooperation costs, which can be aggravated by information asymmetries or ideological preferences.⁷ Hence, neither social group might have an incentive to change an inefficient equilibrium - at least as long as the disenfranchised do not solve their collective action problem and the threat of repression persists. While this rationale does not require the elite to set up inefficient institutions, it presupposes certain changes in the economic conditions in an else passive political environment.

Since these changes are related to groups possessing de facto economic power, this approach basically provides a different, indirect way for an elite to capture democracy and to induce the emergence of inefficient institutions.

The main differences to the model to be discussed are the involvement of the elite in the political process and the temporal setting. Acemoglu et al. (2011) simply assume a more cautious and forward looking elite, who keeps political power right from the transition to democracy, because it is better capable of solving its collective action problem than the poor. Instead of directly using their economic power, they capture the democratic process to shape regulations and to build up entry barriers, which constitutes an explanation for the pro-wealth policy bias of many democracies.

2.3.2. Literature on Inefficient States

The focus of the model on a creation of an inefficient bureaucracy as a result of strategic considerations of a ruling elite in emerging democracy is rather specific, since most of the models concerning inefficient institutions do not include democratic transition. However, it is generally related to the literature on inefficiency and bureaucracy, and more closely to the literature on strategic distortionary actions by politicians creating inefficient outcomes.

First, the model is embedded in the literature on the internal organization of the state (see Acemoglu and Verdier 1998; Dixit 2002; Egorov and Sonin 2011), which does not discuss patronage mechanisms or inefficiency as a means to limit redistribution. Likewise, models focusing on state capacity like Besley and Persson (2009) or Acemoglu (2005) discuss the link between state capacity and inefficiency, but not from a strategic point of view. For instance, in Acemoglu et al. (2013), an inefficient elite-captured democracy arises with high state capacity, because demand for checks and balances eases the elite’s grasp on government.

Second, there are different complementary rationales explaining inefficient states are comple- mentary. Alesina and Drazen (1991) link a war-of-attrition between social groups to delayed reforms, whereas Alder and Ordonez (2012) show how a government using intra-group informa- tion asymmetries can generate rents soaking an elite with the support of the poor in exchange for an inefficient level of redistribution. In a model provided by Chong and Gradstein (2007), income inequality and political bias in favor of the rich lead to less redistribution and lower institutional quality. While these models do not consider emerging democracy, a general framework for elite-captured democracy relying on de facto political power and offsetting de jure changes in political institutions has been discussed by Acemoglu and Robinson (2008). The interplay between de facto and de jure political power is also at work in the model to be discussed, yet it comprises a new focus on strategic choices by the elite that favor the rise of institutional inefficiencies.

A more closely related third set of models lines out inefficiency of redistribution, inefficient bureaucracy and patronage. Different scholars like Becker and Mulligan (2003) have argued that inefficient redistribution may be a means to limit the amount of redistribution (see Coate and Morris 1995; Rodrik 1995; Saint-Paul 1996; Acemoglu and Verdier 2000). In these contributions, however, it remains unclear why partial commitment to the form of redistribution is possible, but not to the amount.

Few models discuss the rise of inefficient bureaucracies for strategic economic reasons. In a similar approach, Stroup and Zissimos (2011) show how an elite may instate a pampered bu- reaucracy to maintain political stability and to reduce competition by keeping an entrepreneurial middle-class from pursuing productive economic activities. Other approaches like Dixit (2010) focus on ineffcient bureaucracies in an autocratic environment. These models hence do not comprise democratic transition, and it remains unclear how a revolutionary threat would be absorbed.

Furthermore, structural approaches to excessive bureaucracies and large public sectors have been brought forward. Besley and Coate (1998) find that inefficiently large governments simply arise due to economic development captured by relative increases in mean income. Limited oversight and costly monitoring are the source of an inefficient bureaucracy in Sarte (1998), whereas Jaimovich and Rud (2009) discuss a model including political bias and institutional quality, where the working class votes for an inefficient public sector. Lastly, time consistency problems and the influence of electoral periods have been shown to increase bureaucracy, since incumbents might want to keep lucrative jobs in case they are not reelected (Haaparanta and Puhakka 1993). None of these models discusses the possibility of a strategic political reasoning behind the setup of an inefficient bureaucracy.

Most closely related to the model discussed in the remainder of the thesis is the literature on politicians distorting policies for strategic reasons. Inefficient policies have been linked to vote-gaining-strategies (Fiorina and Noll 1978; Lizzeri and Persico 2001) or to strategies constraining future politicians (Glazer 1989; Tabellini and Alesina 1990; Aghion and Bolton 1989; Biais and Perotti 2002). Cukierman et al. (1992) emphasize the role of an inefficient tax system in reducing future redistribution, yet there is again no commitment to the level of taxes.

In similar frameworks, Powell (2006) and Hirshleifer et al. (2009) discuss the anticipation of future changes in political power as a source of inefficient policies. They do however not focus on the creation of an inefficient bureaucracy for strategic reasons of non-political groups. Lastly, a model provided by Robinson and Verdier (2002) sort of paves the way for the present approach. It elaborates the effect of inefficient redistribution through jobs in bureaucracy, which are used as a “credible, excludable and reversible method of redistribution that ties the continuation utility of a voter to the political success of a particular politician"(p.1). Yet also this approach does not consider a transition to democracy, but focuses on strategies applied by politicians to stay in power.

The next section will conclude the review and assess differences and weaknesses of the different models in comparison to Acemoglu et al. (2011)

2.3.3. A new Perspective on the Emergence of Inefficient States

By now, the model to be discussed has been embedded into the theoretical framework and the literature on the political economy of inefficient states. Concluding the literature review presented in this section, one can state that most of the related literature neglects strategic interaction in the emergence of inefficient institutions. Inefficient bureaucracies are discussed as a result of organizational problems or linked to rent-seeking governments, but not to the interaction and coalition of social groups. Central commitment problems, e.g. the amount of redistribution or the level of taxes, are not fully explained. In the end, approaches have focussed on existing excessive bureaucracies or patronage, but not on how they emerge in the context of transition to democracy.

The literature has so far tended to focus on proximate causes and beneficiaries of inefficient institutions while neglecting to capture the underlying structures and strategic mechanisms leading to inefficient outcomes. The model provided by Acemoglu et al. (2011) fills this gap which is why it is chosen as a basis for the remainder of this thesis. My analysis and extensions provided in Chapter 4 show that the model indeed allows for broader implications to be inferred. After a short look at related empirical evidence, I will hence start to rebuild the model in Chapter 3.

Since entering into detailed empirical evidence would go beyond the scope of this thesis, I will just provide an overview and a stylized example for the rationale of the model discussed in the next chapter.

2.4. Empirical Evidence

As Acemoglu et al. (2011) note, there are significant differences in corruption of bureaucracy and efficiency of states around the world (World Bank 2004). Furthermore, tax revenues and public expenditures as a share of GDP are also widely dispersed and generally lower in states with inefficient bureaucracies (Acemoglu 2005). For instance, an empirical study on political clientelism by Bustikova and Corduneanu-Huci (2011) underlines the importance of historical state capacity for the reputation of bureaucracy, arguing that a tradition of low state capacity fosters patronage relations. Particularly in emerging democracies, political parties in power are shown to use state resources to build patronage networks. The study concludes that “countries with lower levels of GDP per capita regularly generate budgets that are heavily focused on government consumption" (ibid; p.8), which is consistent with the view of corrupt and low capacity bureaucracies in less-developed countries (see Geddes 1991; Rauch and Evans 2000).

However, such problems are not exclusively linked to emerging democracies. Shleifer and Vishny (1994) analyze the interrelation between government, private and public firms and find broad evidence for excess employment in government firms. For instance, they note that “in Greece all employees and not just top managers of public enterprises turn over when an opposition party wins an election", and that “in many American cities such as Chicago most city jobs are patronage jobs" (ibid; p.996). In the working paper version, Acemoglu et al. (2006) provide some empiric evidence by examining the relationship between bureaucratic wages and public expenditure. They find that when public expenditure is rather low, bureaucratic wages are relatively high, which is in line with their model predictions.

Brazil is quoted as a stylized example of an inefficient state structure, limited public good provision and patronage politics in the distribution of public jobs (see Gay 1990; Weyland 1996; Naritomi et al. 2007: 2012). High inequality and organizational fragmentation foster inefficient structures in general, while in particular the fragmentation of the poor leads to political underrepresentation and hence less redistribution. Bureaucrats and politicians have blocked reforms of the public sector requested by the IMF and international organizations. As Gay (1990) shows, the poor have even supported traditional, elite-attached parties rather than parties with more redistributive agendas. Also, the mechanism of an elite capturing democratic structures in times of democratic transition to secure political power seems to be consistent with Brazil’s development after World War II (see Cohen 1987).

To conclude, it can be noted that some empirical evidence supports the rationale of an inefficient state emerging under patronage politics of a ruling elite. Yet there remains need for further empirical research to test applicability and resilience of models of inefficient states.

[...]

---

¹ See i.a. North (1990); Acemoglu et al. (2004); North et al. (2009); Lizzeri and Persico (2001). 1

² Democratization happens for exogenous reasons and is not modeled as an equilibrium institutional change in this model.

³ Also, the often studied corrupt relationship between private sector and bureaucracy might only constitute a symptom for politically chosen inefficiency.

⁴ For related literature on co-optation, see Bertocchi and Spagat (2001), Gershenson and Grossman (2001) or Balafoutas (2011).

⁵ For an extensive discussion, see Chapter 2.

⁶ Different mechanisms like internal incentives to increase government agencies’ budgets probably play a more important role for the growth of the public sector.

⁷ Facchini and Melki (2013) show in a historical perspective that France had an efficient government size with public spending at 30% of GDP.

¹ I adapt the conceptual framework provided by Acemoglu and Robinson (2006) in Economic origins of Dictatorship and Democracy.

² The incidental institutions view, which describes institutions as a mere secondary product to social interactions, does not fit the present approach either.

³ Concerning the internal organization of bureaucracy, however, the current political economy approach characterizing bureaucracy as a broad apparatus with limited self-interests is adapted (Borcherding 2002, p.6ff).

⁴ The authors claim that this does not cause a loss in economic insight (Acemoglu et al. 2011).

⁵ Evidence from Argentina, Bolivia, Brazil, Peru, Uruguay and the Philippines implies that democratization often happens in recessions, which are associated with lower opportunity costs for the poor(Acemoglu and Robinson 2006). However, the implications of a crisis environment for the emergence of an inefficient state are not considered here.

⁶ Furthermore, the notion of inequality has been discussed. As Bénabou (1996) notes, the relative distribution of earnings and political power is decisive, not inequality per se.

⁷ For instance, Bénabou (2000) proposes a POUM-Hypothesis (possibility of upward mobility).