Principles of comparative politics (Riassunto/Summary), Sintesi del corso di Scienza Politica

Scarica Principles of comparative politics (Riassunto/Summary) e più Sintesi del corso in PDF di Scienza Politica solo su Docsity! PART I. What is comparative politics? 1. WHAT IS COMPARATIVE POLITICS? Political science is the study of politics in a scientific way. Comparative politics is a subfield of political science. 1. Comparative politics is the study of political phenomena that occur predominantly within countries (Politics within nations (1974), Joseph LaPalombara). International politics / international relations is the study of political phenomena that occur predominantly between countries (Politics among nations (1948), Hans Morgenthau). Elections Conflict Party Systems Revolution Foreign Policy Political Economy Executive-Legislative Relations Environmental Politics International » Interest Groups Organizations C, Legislatures 2 e ii International Politics Comparative Politics (Between Nations) (Within Nations) N Comparative politics isthe study of political phenomena in every country except the one inwhichthe student resides. /n Nigeria, the study of comparative politics would not include the study of Nigerian politics as it would be treated as a separate subfield of political science. 3. Comparative politics is the study of political phenomena using the method of comparison, which consists in seeking answers to questions about politics by comparing and contrasting different polities. Comparison is the essence of every scientific field, including political science, because the method of science isthe method of comparison. This definition would make ‘comparative politics’ a synonym of ‘political science’. 2. WHAT IS SCIENCE? Science is a method for provisionally understanding the world: the best answer is not the final answer, it is just the best so far. Science is a quest for knowledge that relies on self-criticism. Scientific statements must be falsifiable: it does not mean thata scientific statement will ever be falsified, butthere must be a possibilitythat it could be falsified. Falsifiable statements are potentially testable: all that is required isthat we can conceive of a way to test it even if we do not currently have the equipment to test it, therefore, it does not have to have been tested to be scientific. Sorts of statements that are not scientific: ® tautologies, because they are true by definition. Triangles have three sides: it is impossible to observe a triangle that does not have three sides because, by definition, a figure that has three sides is a triangle. It is not falsifiable, hence, it is not scientific; e statements about unobservable phenomena. God created the world, God exists. It cannot be tested, hence, it is not scientific. This does not mean that nonscience is nonsense or that these claims are necessarily false. It only means that these statements are unscientific. 2.1 The scientific method: falsificationism The scientific method describes the process by which scientists learn about the world. Although there is no scientific method clearly written down that is followed by all scientists, it is possible to characterize the basic features of the scientific method: 1 N question / puzzle + we observe the world and come up with a question. How did a country become a democracy? theory / model / explanation + a set of logically consistent statements that answer the question: they tell us why the things we observe occur. Why do we have to tell why the things we observe occur? The principle of uniformity of nature asserts that if X causes Y today, then it will also cause Y tomorrow and the next day and so on; if it does not, then X is not a cause. This principle does not assert that nature is unchanging, only that the laws of nature do not change. If this principle is rejected, we must accept the possibility that things just happen for no reason. How do we construct a theory? When generating a theory, it is useful to think of the starting puzzle (the question) asthe end result of some previously unknown process. This ‘unknown process’ becomes our model that explains the puzzle. A good model contains only what is needed to explain the phenomenon that puzzles us and nothing else. Details are important only ifthey are crucialto what we are trying to explain, otherwise they can only distract our attention from the question. For this reason, we should evaluate models in terms of how useful they are for achieving your goal. Think of models as similar to subway maps: subway maps are simplified pictures of the city and they are geographically inaccurate, but it is not important because they are designed only to help you move efficiently around the city using the subway; implications / hypothesis + we deduce them from the model otherthan those that we initially set out to explain. We ask: if the prior world that we created to explain the phenomenon that we originally found puzzling really did exist, what else should exist / what else should we be able to observe? Good models are those that: e produce many different implications + each implication represents another opportunity for the model to fail. However, what really matters is the severity and the variety of implications: we gain more confidence if the model has survived a few harsh implicationsthan if it has survived many easy ones; e produce surprising implications + they tell us something that we would not know in the absence of the model, especially when the model tells us something that we already know. However, surprise is best appreciated in small doses, otherwise it may seem like the model or what we know about the world is wrong. observe the world / test hypothesis + to examine whetherthe implications ofthe model are consistent with the observation. The standard practice consists in asking if other different models might also explain the phenomenon of interest. When this is the case, we compare the implications for the same phenomenon, and we will most 2 General form —Example Specific example 7 P, then Q Hour theory T'is correct, then we If our theory is correct, then we should should observe some implication /. observe that rich democracies live longer than poor democracies. valid Not Q We do not observe implication /. Rich democracies do not live minor premise longer than poor democracies. Therefore, not P. Therefore, our theory T's incorrect. Therefore, our theory is incorrect. There are two ways of testing theories: 1. deductive approach + theory precedes observation; 2. inductive approach + observation precedes theory: observations of the world are collected and then used to develop theories about why certain political phenomena occur /to identify the causes of political events. A cause is a necessary or sufficient condition: e necessary condition + a circumstance in whose absence the event in question cannot occur: Y never happens unless X happens / if no X, no Y. Oxygen (cause, X) is a necessary condition for fire (effect, Y); 7° N \ ( Ettect E Present n % 14 e sufficient condition + a circumstance inwhose presence the event in question must occur: Y always happens if X happens / if X, then Y. Fire (cause, X) is a sufficient condition for smoke (effect, Y). Feffect E Present | Cause Cs Present ) \ / Induction is problematic because it relies on affirming the conseguent and the theory is never exposed to potential falsificationism. The most common method of induction in comparative politics is the comparative method, also known as “John Stuart Mill’s methods’. In A system of logic (1872), Mill outlined: the method of agreement + it compares cases that agree in regard to the political phenomenon to be explained. Suppose that we want to explain what causes democracy. We observe two or more cases that are democracies / that agree in regard to the political phenomenon. Ethnically Parliamentary Country Democracy Wealth homogeneous system UK Yes Yes Yes Yes Belgium Yes Yes No Yes US Yes Yes Yes No We observe that a) ethnic homogeneity and having a parliamentary system are not necessary conditions (Belgium, the USA), b) wealth is a potential necessary condition (democracy is caused by wealth). ii. the methodof difference + it compares cases that differ in regard to the political phenomenon to be explained. Suppose that we want to know whether wealth is a sufficient condition for democracy. We observe two or more cases that are non-democracies / that differ in regard to the political phenomenon. Ethnically Parliamentary Country Democracy Wealth homogeneous system UK Yes Yes Yes Yes Belgium Yes Yes No Yes Us Yes Yes Yes No Mexico (priorto 1990) No Yes Yes No We observe that a) wealth is not a sufficient condition for democracy (Mexico is wealthy, but not a democracy), b) ethnic homogeneity is neither necessary (Belgium) nor sufficient (Mexico), c) having a parliamentary system is not necessary (the USA) but it may be sufficient (Belgium, the UK). 2.3 False myths about science 3. Science proves things and leads to certain and verifiable truth > false. The best science can offer are tentative statements. We are more confident about what we do not know than what we do know: the process of scientific accumulation can be thought of as the evolution of our ignorance. Science can be done only when experimental manipulation is possible + false. For theories to be scientific, they only need to be falsifiable. There is no need for them to be carried out in an experimental setting, like a lab. Scientists are value-neutral > false. Science is value-neutral, but scientists may not be value-neutral. The types of research questions that are asked and the interpretation of scientific results are likely to be infused with the specific values and biases held by scientists and those who use their research. The lack of diversity in most disciplines (gender, race, income, sexuality, religion, ethnicity, etc.) means that certain research areas are less studied and that certain viewpoints are excluded or less privileged. We should try to conduct our studies in such a way that someone who does not share our biases can determine if our arguments and evidence are reasonable. Politics cannot be studied in a scientific manner > false. WHAT IS POLITICS? Politics comprises the subset of human behaviour that involves the use of power. Based on the assumption that we all have goals, power is involved whenever individuals cannot accomplish their goals without eithertrying to influence the behaviour of others or avoiding the influence exerted by others. Game theory is not a theory as in the way we studied it before, but a fundamental tool that political scientists use for analysing situations in which the choices of one player depend on the choices made other players. 3.1 The Exit, Voice and Loyalty (EVL) game We use Albert Hirshman's Exit, Voice, and Loyalty (1970) to understand the central characteristics of the power relationship between the government and the citizen (the players). The citizen has three possible choices: e exit + you accept the negative change in your environment and alter your behaviour to achieve the best outcome possible from this change; e use voice + complaining, protesting, lobbying or taking other forms of direct action to try to change the environment back to its original condition; e demonstrate loyalty + you accept the negative change in your environment and make no change to your pre-existing behaviour. Stimulus Exit Voice Loyalty Your state increases Reallocate portfolio Organize tax revolt Pay taxes, keep your taxes. to avoid tax mouth shut increase When is one's behaviour / decision political? When you start considering whether to respond with exit, voice or loyalty, and when you choose the voice option (choosing the exit and loyalty options is not political). Games have rules about how decisions are made. The basic rule is that players choose to do what they believe is in their best interest. The interests of the players are reflected in the payoffs / benefits associated with each possible outcome of the game. Players prefer outcomes with higher payoffs to outcomes with lower payoffs. The prehistory / background of the EVL game is that the government has caused a negative change in the environment of the citizen that resulted in a transfer of some payoff from the former to the latter. The government has introduced a tax hike leading to an increase in revenue for the government and less income for the citizen. To represent the EVL game we use an extensive form game, which employs a game tree that allows us to see what happens when the players take turns (sequentially) to make decisions. A game tree consists of choice nodes linked in a sequence. Choice nodes are points at which a player must choose an action. Each choice node is marked by the name of the player making the choice at that point in the game. The first choice node in a game / the place where the game starts is called ‘initial choice node’. The choice nodes are linked to other choice nodes by lines called ‘branches’. The branches represent the actions that can be taken at each choice node. A branch that is not followed by another choice node leads to one of the potential outcomes of the game. Each ofthe potential outcomes of the game is labelled with the payoffs earned by each player inthat outcome. Citizen initial choice node Exit Loyalty ‘01: Government keeps benefit; Voice 02: Government keeps benefit; «itizen opts for some substitute | citizen suffers loss Government Respond Ignore 03: Government returns benefit to citizen Citizen 04: Government keeps benefit; ©5: Government keeps citizen opts for some substitute benefit; citizen suffers loss The citizen’s payoffs are shown first because he is the first to make a choice; the government’s payoffs are shown second: We assume that E <0, therefore, 0>0—c, the citizen will choose to remain loyal. Scenario 3 > the citizen has a credible exit threat (E > 0), the state is autonomous (L < 1). Citizen Exit Loyalty The expected outcome of the game is that the citizen exits from the beginning and the government gets to keep the Voice 0,141 benefit; SormnenE The citizen receives E and the government receives 1 (E, 1); Respond I pon — The SPE is (exit, exit; ignore). Citizen Exit Loyalty E-61 0-G14L The final choice node of the game has the citizen deciding whether to exit or remain loyal: a) if the citizen chooses to exit, he will receive a payoff of E — c, b) if the citizen chooses to remain loyal, he will receive a payoff of 0—c. We assume that E>0 andc>0, therefore, E-c>0— c, the citizen will choose to exit. The choice node prior to the final choice node has the government deciding whether to respond (positively) or ignore: a) ifthe government chooses to respond, it will receive a payoff of L, b) ifthe government chooses to ignore, the citizen will choose to exit at the final node and its payoff will be 1. We assume that L < 1, therefore, 1 > L, the government will choose to ignore. The initial node hasthe citizen deciding whetherto exit, remain loyal or use his voice: a) ifthe citizen chooses to exit, he will receive a payoff of E, b) ifthe citizen chooses to remain loyal, he will receive a payoff of 0, c) ifthe citizen chooses to use his voice, the government will choose to ignore, then the citizen will choose to exit and his payoff will be E — c. We assume that E > 0, therefore, E > E— c, the citizen will choose exit. Scenario 4 > the citizen has no credible threat (E < 0), the state is autonomous (L< 1). Citizen Exit Loyalty I Voice 0,141 | gets to keep the benefit. The expected outcome of the game is that the citizen remains loyal from the beginning and the government Government The citizen receives 0 andthe government receives 1 + Respond Ignore L(0,14+L1). 1-gL The SPE is (loyalty, loyalty; ignore). Citizen | Exit Loyalty E-G1 0-c14L 10 The final choice node of the game has the citizen deciding whether to exit or remain loyal: a) if the citizen chooses to exit, he will receive a payoff of E — c, b) if the citizen chooses to remain loyal, he will receive a payoff of 0—c. We assume that E <0andc>O0, therefore, 0—c>E—c, the citizen will choose to remain loyal. The choice node priorto the final choice node has the government deciding whether to respond (positively) or ignore: a) ifthe government chooses to respond, it will receive a payoff of L, b) if the government chooses to ignore, the citizen will choose to remain loyal at the final node and its payoff will be 1 + L. We assume that L < 1, therefore, 1 > L, the government will choose to ignore. The initial node hasthe citizen deciding whether to exit, remain loyal or use his voice: a) ifthe citizen chooses to exit, he will receive a payoff of E, b) if the citizen chooses to remain loyal, he will receive a payoff of 0, c) ifthe citizen chooses to use his voice, the government will choose to ignore, then the citizen will choose to remain loyal and his payoff will be 0—c. We assume that E <0, therefore 0>0 —c, the citizen will choose to remain loyal. Summary of the SPE with their expected outcomes: The state The citizen Is Autonomous Is Dependent (L<1) (£>1) Has a Credible Exit Threat (Exit, Exit; Ignore) (Voice, Exit; Respond) (E> 0) Outcome 1 Outcome 3 Has no Credible Exit Threat (Loyalty, Loyalty; Ignore) (Loyalty, Loyalty; Ignore) (E<0) Outcome 2 Outcome 2 3.2 Conclusion About the power relationship between citizen and government, we conclude that: autonomous states never back down; the government will be willingto respond positivelyto the citizen only when a) the citizen has a credible exit threat (E > 0) (exit, voice), and b) the government is dependent on the citizen (L > 1). If the government attempts to take a benefit away from a citizen with a credible exit threat, the citizen will use his voice, and because the state depends on him, it will respond positively to his demands. Because the government knows that it will eventually have to respond positively, it will choose not to seize the citizen’s resources in the first place. As a result, the citizen will never have to use his voice because the state will always do whatever this type of citizen wants. According to the structural Marxist view of the state, even though the government is dependent equally on both labour and capital, capitalists exercise tremendous power over the government than workers because they possess credible exit threats. It is also the case of the US government during the 2008 global financial crisis, who moved quickly and enthusiastically to bail out rich New York banks (on October 3, Bush signed the Troubled Asset Relief Program (TARP), which authorized the Treasury to spend up to $700 billion to help banks) while moving more slowly and with greater reluctance to come to the aid of troubled car manufacturers (in December, Bush authorized that $13.4 billion of the TARP money would be made available to auto manufacturers). The state was dependent on the support of both the financial sector and the automobile manufacturers; however, the 11 financial sector had a credible exit threat (it can easily move around its assets), whilethe manufacturers could not credibly commit to exiting, at least in the short-term (automakers' assets are fixed). Sufficiently powerful people do not need to use their voice because they are already getting other people to do what they want. In many cases, the most powerful are precisely those citizens who are least likely to take action or use their voice, so the political scientist will never be able to observe them using their power. For this reason, it might be misleading to infer that citizens who do not use their voice have no significant influence over government policy. /t would be wrong to infer that the President of the Republic lacks power because he rarely uses his right to veto legislation. His right to veto is enough to ensure that legislatures send only the bills that the president wants; when the citizen has no credible exit threat: o the government can choose to ignore and take away the citizen’s benefits, and the citizen can only accept the situation. /n the USA, the Democrats depend on African American voters, but African American voters do not have a credible exit threat; o itis not so easy to learn whether a government is dependent or autonomous because both types of government will simply ignore this type of citizen. Therefore, if we see a state ignoring a citizen, we should not inferthat the state is necessarily autonomous: the state may be reliant on the citizen but can safely ignore her if she does not have a credible exit option (China, Iran, North Korea, East Germany). there are also things that the EVL game does not explain: citizens use their voice only when they expect it to be effective, then why do we see governments being unresponsive to the demand of their citizens? Maybe the citizens do not see protesting as costly: they enjoy using their voice even if they know it will not be successful, to network with other like-minded people, etc. Maybe there is incomplete information: the state does not know if the citizen has a credible exit threat or not, the citizen does not know if the state is dependent or not, etc. For each possible circumstance, each player chooses his best reply, the best choice in response to each of the possible choices that the other player might make, the choice with the highest payoff. The best replies for player A are: e if player B refrains > if player A refrains, he will get a payoff of 3, and if he steals, he will get a payoff of 4. Therefore, player A's best reply is to steal. We indicate this by placing a line under the 4; ® if player Bsteals > if player A refrains, he will get a payoff of 1, and if he steals, he will get a payoff of 2. Therefore, player A's best reply is to steal. We indicate this by placing a line under the 2. The best replies for player B are: e if player A refrains + if player B refrains, he will get a payoff of 3, and if he steals, he will get a payoff of 4. Therefore, player B's best reply is to steal. We indicate this by drawing a circle around the 4; e if player Asteals > if player B refrains, he will get a payoff of 1, and if he steals, he will get a payoff of 2. Therefore, player B's best reply is to steal. We indicate this by drawing a circle around the 2. To locate the NE, we need to find the cell / strategy combination in the payoff table in which both equilibrium strategies marked as best replies. In the state of nature game, there is only one NE, but there is no rule that there will always be one NE. B Refrain Steal The expected outcome of the game is that both players steal. Refraîn 3,3 1® Each player receives a payoff of 2. i The strategy combination that represents the NE in the state of nature game is (steal; steal). Steal 4,1 2 pi We say that a player has a dominant strategy if that strategy is a best reply to all of the other player's strategies: here, the dominant strategy for each player against the other is to steal. A dominant-strategy Nash equilibrium occurs when both players have a dominant strategy. Itis oddthat the NE outcome from the game happens to be the second worst outcome for both players. The absence of cooperation represents a sort of dilemma: individual rationality leads actors to an outcome that is inferior. Both players could be better off ifthey chose to refrain: they would both get a payoff of 3 (refrain) instead of the 2 (steal). Imagine that player A and player B promised not to steal from each other because this would make them both better off. Do you thinkthat they would feel comforted by such promises? No, because: e ifyouarethe only one whois stickingto your promise and your opponent is the one who breaks his promise and startsto steal, your best response is to stop refraining and start stealing as well + you will increase your payoff from 1 (refrain) to 2 (steal); ® ifyouknowforsure that you opponent is refraining, your best response is to stop refrainingand start stealing + you will increase your payoff from 3 (refrain) to 4 (steal). 15 Thus, simply promising to stop stealing because it is mutually destructive is not sufficient to actually stop the players from stealing. The civil society Hobbes's solution to this problem was to create “some coercive power to compel men equally to the performance of their covenants, by the terror of some punishment greater than the benefit they expect by breach of their covenant” + a sovereign with sufficient power created by an implicit social contract between individuals inthe state of nature: individuals would give up their natural rights in exchange for civil rights that would be protected by the sovereign. A natural right is a universal right that is innerent in the nature of living beings (given bythe nature), as such, it exists in the state of nature. A civil right does not arise naturally but is instead created by the state (through laws), as such, it cannot exist inthe state of nature. Hobbes believed that life in the state of nature was sufficiently bad that individuals would and should be willing to transfer everything they had to the sovereign in exchange for protection. Other social contract theorists, like Rousseau and Locke, believed that individuals in the state of nature should find ways to use a third party, such as a sovereign, to achieve only a limited degree of cooperation. Although there are differences between them, the contractarian view of the state sees the creation of the state, resulting from a social contract between individuals in the state of nature, as a third-party enforcer who can punish individuals who engage in socially destructive behaviour that violates the social contract. The result is a civil society in which individuals live with a state. The civil society game The state’s punishments are structured in such a way that ‘steal’ would no longer be a dominant strategy for individuals. But is it sufficient? It depends. We can see what it depends on by solving the civil society game. In the civil society, in addition to player A and B, we also have one passive player in the background, the state, who has sufficient physical force to punish people who choose to steal rather than refrain. The punishment will be of value p. The four outcomes are the same as before: both players refrain, both players steal, player A steals and player B refrains, player B steals and player A refrains. B Refrain Steal We give from 4 to 1 from the most preferred to the least preferred payoff. These payoffs are Refrain 3,3 1,4-p called ‘cardinal payoffs' because they tell us how much more a player values one outcome A e el l]2eee-—2—- I compared with another. Steal Solving the civil society game Now that we know the players, the choices available to them, and how they value each possible outcome, we solve the civil society game for the Nash equilibrium (NE) as before. 16 The best replies for player A are: e if player B refrains > if player A refrains, he will get a payoff of 3, and if he steals, he will get a payoff of 4 — p.3>4- p, as long as p> 1 (biggerthan the difference between 4 and 3), therefore, player A's best reply is to refrain. We indicate this by placing a line underthe 3; ® if player B steals > if player A refrains, hewill get a payoff of 1, and if he steals, he will get a payoff of 2 — p. 1>2-p, aslongasp>1, therefore, player A's best reply isto refrain. We indicate this by placing a line under the 1. The best replies for player B are: e if player A refrains + if player B refrains, he will get a payoff of 3, and if he steals, he will get a payoff of 4— p.3>4-p, aslong as p> 1, therefore, player B's best reply is to refrain. We indicate this by drawing a circle around the 3; e if player Asteals > if player B refrains, he will get a payoff of 1, and if he steals, he will get a payoff of 2—p. 1>2- p, aslong as p> 1, therefore, player B's best reply is to refrain. We indicate this by drawing a circle around the 1. In the civil society game, there is only one NE, but there is no rulethat there will always be one NE. B . The expected outcome of the game is that both Refrain Steal players refrain. Refrain 3® 14-p Each player receives a payoff of 3. The strategy combination that represents the NE in the civil society game is (refrain; refrain). Steal 4-p@ 2-p,2-p x The dominant strategy for each player against the other isto refrain: as long as the punishment level imposed by the state is sufficiently high, players will refrain no matterwhat the other player decides to do. Although it seems that by creating a state we can get individuals to give up the sorts of behaviour that made life in the state of nature “solitary, poor, nasty, brutish, and short”, problems are not all solved. Why would anyone want to be the sovereign and why would he be willing to do us all a favour by acting as our ‘policeman’? Let us portray the citizens as being engaged in an exchange relationship with the sovereign: the sovereign agrees to police us in exchange for taxes of value t that citizens pay. However, the citizen's choice to leave the state of nature forthe civil society will depend on the level of taxation imposed by the state: the taxation rate charged by the state must not be so large that individuals preferthe state of nature to the civil society. The civil society game illustrates that the state will impose a tax of size t on the citizens for allowing them to live ina civil society. We indicate this by subtracting t from the payoffs of each player in each cell. Note that because the citizen must pay the tax in every cell of the game, the expected outcome of the game does not change: both players will still choose to refrain. We compare the state of nature game and the civil society game: 17 4.5 Can cooperation occur without the state? Some scholars have argued that cooperation can emerge through a decentralized process in the state of nature as long asthe individuals inthe state of nature repeatedly interact with each other and care sufficiently about the future benefits of cooperation. We saw that cooperation or refraining was not possible when the state of nature game was played once. Let us examine what happens if the individuals in the state of nature play the game over and over again. First, we introduce two new concepts: e the discount factor + it tells us how much people care about the future, specifically, how much future benefits are discounted compared with today's benefits. Low discount factors mean that people do not value the future very much: they value the benefits they receive today much more than the benefits they will receive tomorrow. High discount factors mean that people value the future a lot: they value the benefits they will receive tomorrow almost as much as the benefits they receive today. The higher the discount factor, the more you care about the future; the lower the discount factor, the less you care about the future. Imagine that you had a choice of receiving $1,000 today or $1,000 in a month/'s time: o if receivingthe money today or in a month's time did not matter to you, your discount factor d would be 1 + future payoffs are worth as much to you as today's payoffs; o if receivingthe money in a month's time was worth something to you, but not as much as getting it today, your discount factor would be 0 < d <1 > future payoffs are worth something to you; o if receiving the money in a month's time was worthless to you, your discount factor would be 04> future payoffs are worth nothing to you. e the present value of some stream of payoffs > imagine that your parents promise to give you $1 every day from now into the future. How much all the dollars that you expect to receive in the future worth to you today /what isthe present value of this stream of future payments / payoffs? © The $1 you receive today is obviously worth $1 to you today. o The S1 you expect to get tomorrow is worth only $1d today. o The $1 you expect to receive in two days time is worth only $1d 2 (it will be discounted by your discount factor a second time). o The $1 you expect to get in three days’ time is worth only $1d5. Present value (promise) = 1 + 1d + 1d ? + 145 + 1d*+ 14% +...+ 1d°° Present value (promise)= 1+d+d?+d3+d4+d5+...+d° Present value (promise)=1/(1—d) If your parents had promised you $5 every day indefinitely into the future, the present value of the promise would be 5 / (1-d). If they had promised you $3, the present value of the promise would be 3 /(1-d), and so on. Now that players A and B have to play the state of nature game over and over again, they have to decide how to play the game in each round (or period). One strategy that the players might employ is the grim trigger strategy, which 20 says that a player will refrain (cooperate) as long as the other player refrains; but if the other player steals instead, the first player will steal from him in the next round and in all future rounds. How much will the players get if they both choose to refrain? Both players will get a payoff of 3 in every period. This payoff will be discounted by the discount factor d every time the game is played. Present value (refrain) = 3 + 3d + 3d ? + 3d 3 + 3d* + 3d5 +... + 3d°° Present value (refrain)=3+d+d?+d3+d4+d5+...+ d°° Present value (refrain) =3/(1—d) How much would player A get if he changed his strategy from refrain to steal while player B continued to refrain (no NE)? Player A would get a payoff of 4 in the first period, while player B continues to get a payoff of 3. If player B decides to use the grim trigger strategy, he will respond to player As stealing by stealing in all the future rounds of the game. Given that player B is always going to steal, the best that player A can do now is to continue stealing. As a result, player A will get a payoff of 4 in the first period but a payoff of only 2 discounted by the discount factor d in every period thereafter. Thus, player A's present value of unilaterally changing his strategy from refrain to steal is: Present value (steal) = 4 + 2d + 2d 2 + 2d 3 + 2d4* + 245 +... + 2d° Present value (steal) =4+2d(1+d+d?+d*+d4+d5+...+d°) Present value (steal)= 4+2d/(1-—d) Player A will not unilaterally choose to change his strategy to steal ifthe present value of refraining is greater than the present value of stealing (‘Present value (refrain) > Present value (steal) ), i#3/(1—d)is greaterthan4+2d/(1 — d). Player Bwill also preferto refrain ratherthan steal because he has the same payoffs as player A. 3/(1-d)>4+2d/(1-d) 3>4+2d -2d<1 2d>1 d>1/2 d>0.5 This also means that the state is not strictly necessary for cooperation between individuals in the state of nature. However, even if it is not strictly necessary, not having a state may not be the best thing: e aslongas individuals are sufficiently concerned about the potential benefits of future cooperation, ‘refrain’ can be sustained asa NE in the infinitely repeated state of nature game when both players use a grimtrigger strategy and the discount factor is greater than a half; e however, ‘steal’ can also be sustained as a NE in the infinitely repeated state of nature game when both players use a grim trigger strategy and the discount factor is greaterthan a half: if your opponent is always goingto steal, then you never have an incentive to unilaterally change your strategy, you will always steal as well. Game theory cannot tell us which equilibrium is most likely to occur in these circumstances. As a result, there is no reason to believe that the cooperative outcome will be any more likely to occur than any of the other equilibrium outcomes. Moreover, it actually takes a lot of effort for the players to sustain cooperation in the state of nature because everyone has to monitor everyone else to see who is stealing and who is not. It also requires that the individuals get together to punish those people who have been caught stealing. 5. DEMOCRACY AND DICTATORSHIP: CONCEPTUALIZATION AND MEASUREMENT 5.1 States, governments, and regimes Astate is an entitythat uses coercion and the threat of force to rule a given territory. With the exception of relatively rare cases in which states merge and break up (the formation of the United Arab Emirates in 1971, South Sudan in 2011), contemporary states are generally quite stable. A government is a set of people who run the state or have the authority to act on behalf of the state at a particular point in time: they arethe meansthrough which state power is exercised. While states are generally stable overtime, governments come and go. A regime is a set of rules, norms, or institutions that determine how the government is constituted and organized, and how major decisions are made. They are generally classified as being either democratic or dictatorial; democratic and dictatorial regimes also come in different types (presidential, parliamentary, or semi-presidential democratic regimes; dictatorial regimes run by the military, a monarchy, or civilians). 5.2 Democracy and dictatorship in historical perspective Democracy Democracy has come to be considered an important and desirable political system with a highly positive connotation only relatively recently during the second half of the 20°" century. Even countries widely considered to be dictatorships have professed their support for democracy and have adjusted their definition of democracy so that the word could be applied to theirform of regime. Some have even made reference to ‘people’ or ‘democracy’ in their names: the Democratic People's Republic of Korea, the Democratic Republic of Congo, the Lao People's Democratic Republic. These countries even adopt seemingly democratic institutions, such as elections, legislatures, and political parties. However, historically, democracy was viewed as an obsolete, dangerous, and unstable political system. The Greek word ‘demokratia’ is often translated as ‘rule by the people’, but it refers more specificallyto the ‘common people’, people with little or no economic independence / poor and politically uneducated. “Everybody [...] knew that democracy, in its original sense of rule by the people or government in accordance with the will of the bulk of the people, would be a bad thing” (C. B. Macpherson). e One reason was that the debates concerning the appropriate form that regimes should take typically occurred among elites rather than the common people. The earliest debates can be dated to around 520 B.C. in Persia. However, it was perhaps Plato and Aristotle the first to think systematically about the different forms that regimes could take: o The Republic, Plato + he believed that allowing the common people to rule in a democracy would result in ochlocracy or mob rule and would lead to short-lived democracies in which people could quickly surrender to a tyrant. The political decision making should be based on expertise, only trained statesmen should guide the state; o Politics, Aristotle + he also believed that democracy would be the most dangerous form of regime because it is characterized by class rule, in which the common people govern for themselves rather than for all. n n Polyarchies (Ideal Type) Apartheid Liechtenstein pre-1984 US today South Africa Switzerland pre-1971 S US pre-1830 France pre-1945 È po n 8 3 Ò China Soviet Union Inclusion Countries in the top left have a high level of contestation and a low level of inclusion (Apartheid South Africa and US pre-1830: the right to vote was still not expanded enough); countries in the bottom left have low levels of both contestation and inclusion (China: there is only one party and there are no elections above the municipal level); countries in the top centre have a high level of contestation and a moderate level of inclusion (Liechtenstein pre- 1984, Switzerland pre-1971 and France pre-1945: they had multiparty elections but universal suffrage was applied only to men); countries in the bottom right have a low level of contestation and a high level of inclusion (the Soviet Union: everyone was allowed to vote and participate, but there was only one political party). Countries in the top right have high levels of both contestation and inclusion (most countries that we immediately recognize as being democracies today) and are called ‘polyarchies’ by Dahl: he did not believe that any country exhibited, or could ever exhibit, sufficient levels of contestation or inclusion to rightfully be considered a true democracy. Dahl's measure is a continuous measure with intermediate categories within a given range / ittreats regime type as a continuum: there can be countries which are more or less democratic / equally democratic and dictatorial. The opposite is the dichotomous measure with only two categories, treating regime type as a dichotomy: countries are either a democracy or a dictatorship. Three measures of democracy and dictatorship Although there are several different measures of democracy and dictatorship, we focus here on three that are commonly used in the comparative politics literature: 1. the Democracy-Dictatorship (DD) measure (from Cheibub, Gandhi, and Vreeland) + it provides an annual measure of democracy and dictatorship for 199 countries from 15 January 1946 to 31° December 2008. According to the conceptualization of democracy underlying the DD measure, democracies are “regimes in which governmental offices are filled as a consequence of contested elections.” So, this conceptualization of democracy has two primary dimensions / for a regime to be considered democratic: governmental offices must be elected + a) the chief executive office, and b) the legislature; ii. there must be contestation: c) there is more than one party competing in the elections + for a contested election to take place, voters must have at least two alternatives to choose from. Elections in which only one party competes or in which voters are confronted with a single party list are therefore not considered contested: Syria, Turkmenistan, Vietnam. Therefore, the factthat there are elections is not sufficient to distinguish between democracies and dictatorships because virtually every country in the world has held legislative or presidential elections at one time or another and would have to be considered democratic by this standard; d) alternation in power under identical electoral rules has taken place + the individual who is the chief executive is replaced through the electoral process by someone else. An alternation occurred in Mexico when Enrique Pefia Nieto of the Institutional Revolutionary Party (PRI) took over the presidency from Felipe Calderén of the National Action Party (PAN) in 2012. If the incumbent ruler does not demonstrate that he is willing to give up power after losing an election, thenwe have no way of trulyknowing whetherthe country is a dictatorship or a democracy. For example, it is impossible to distinguish between regimes in which the incumbents are always in power because they are popular (but would give up power if they lost) and those in which the incumbents hold elections only because they know they will not lose: these two scenarios are observationally equivalent. In each of these countries, multiparty elections have been held; however, a single party has been in power for long periods of time: e until the incumbents lost in Japan (Liberal Democratic Party, LDP) and Mexico (PRI) and willingly gave up power, it was unclear whether these countries should be considered democracies or dictatorships; ® in Malaysia, three multiparty elections were held between 1957 and 1969. The incumbent government (Alliance Party coalition) won all three elections. However, on the third election, the Alliance won with a reduced majority. The unease and anxiety ledto declaring a state of emergency, closing parliament, and rewriting the constitution such that it (and its successor, the National Front coalition) never lost an election. That the incumbent government was unwilling to give up power when it ‘lost’ the third election suggests that it would have been unwilling to give up power if it had lost either of the two previous ones; e the incumbent government of Botswana (Botswana Democratic Party) has never lost an election since the country became independent in 1966. According to the DD measure, Botswana is coded as a dictatorship precisely because it is impossible to know for certain whether the incumbent government would willingly give up power if it lost. It is considered preferable to mistake a democracy for a dictatorship rather than mistake a dictatorship for a democracy. Contestation involves three elements: e ex ante uncertainty > the outcome of the election is unknown before it happens. /t rules out those countries in which there is an absolute certainty as to which candidate or party is going to win before the voters go to the polls (Iraq under Saddam Hussein); e ex postirreversibility > the winner ofthe election actually takes office. /t rules out countries like Algeria in 1991, where the army intervened to prevent the Islamic Salvation Front from taking office following its success in the first round of legislative elections; e repeatability > elections that meet the first two criteria must occur at regular and known intervals. /t rules out countries like Weimar Germany in the 1930s when the Nazi Party came to power through democratic elections but then cancelled further electoral contests. It completely ignores inclusion because countries around the world exhibited almost no variation in their level of inclusion forthe studied time frame 1946 — 2008 (by 1946, all but a handful of countries had adopted universal suffrage for their elections). 26 N the Polity IV measure (from Marshall, Gurr, and Jaggers) + it provides an annual measure of democracy and autocracy for 167 countries from 1800to the present. The Democracy and Autocracy scores for each country both range from O to 10. From these two measures, a Polity Score for each country is calculated as the Democracy Score minus the Autocracy Score. As a result, the Polity Score for each country ranges from a minimum of -10 (as autocratic or dictatorial as possible) to a maximum of 10 (as democratic as possible): e democracies, if their Polity Score is from +6 to +10; ® anocracies or mixed regimes, if the Polity Score is between —5 and +5; e dictatorships, if their Polity Score isfrom £ to-10. A country's Polity Score is based on five different dimensions, each worth a different number of points: the competitiveness of executive recruitment; ii. the openness of executive recruitment; democratic governments must ili. the constraints that exist onthe executive 7 be limited govermments iv. the regulation of political participation. The possible score is: Contribution to Contribution to Contribution to Democracy Score Autocracy Score Polity Score Regulated 0 o 0 Multiple Identity 0 o 0 Sectarian 0 1 A Restricted 0 2 -2 Unregulated 0 o 0 v.. the competitiveness of political participation. The possible score is: Contribution to Contribution to Contribution to Democracy Score Autocracy Score Polity Score Competitive 3 0 3 Transitional 2 0 2 Factional 1 0 1 Restricted 0 1 -1 Suppressed 0 d 2 the Freedom House measure + it provides an annual measure of global freedom for countries around the world since 1972. The 2015 Freedom in the World survey covers 195 countries and 15 related and disputed territories. Although freedom is not technically a measure of democracy, many scholars use it as if it were, presumably under the assumption that democracy and freedom are synonymous. It is up to you to decide whether it is reasonable to assume that the more freedom exhibited by a country, the more democratic it is. A country's Freedom House score is based on two dimensions: a country's level of political rights + measured by a series of 10 questions, each worth between 0 and 4 points: 0 indicates the lowest level of freedom and 4 the highest level of freedom (there are two additional discretionary questions that can add or subtract up to 4 points; the maximum score 27 — the study of democratic transitions would suggest that a dichotomous conceptualization is more appropriate because a transition seems to imply movement from one distinct regime to another; — the study of the effect of foreign intervention on a country's level of democracy would suggest that a continuous conceptualization is more appropriate. e validity > it refers to the extent to which our measures correspond to the concepts that they are intended to reflect: does our indicator actually measure the thing that it is supposed to measure? Several issues arise when we think about validity. Here, we briefly considerthree: o attributes (that make upthe abstract concept) + there are no rules for determining which attributes must be included when measuring a particular concept, but there are certain issues that scholars can take into account: — having too many attributes, a particular concern with substantive measures, because it reduces the usefulness of the measure. For example, Dahl was worried that there would be no country in the world that could be classified as a true democracy if too many attributes were included; — havingtoofew attributes, a particular concern with minimalist measures, because we would not really capture all of what we think of when we think of democracy. For example, scholars criticize the DD measure for not taking into account Dahl's notion of inclusion. o aggregation issues + when the researcher has decided upon the attributes that constitute the abstract under consideration, he has to decide how to combine these attributes into a single measure of concept. The DD measure does not require aggregation rules because a country is either a democracy or a dictatorship based on whether it passes a set of necessary and sufficient conditions. Polity IV and Freedom House require aggregation rules: none of the individual attributes is explicitly necessary for a country to be considered democratic because a country is classified as democratic whenever it scores a high enough number of points. You should only ask if the aggregation rules employed by them are appropriate and justified. For example: — isitappropriatefor Freedom House to weight each of the attributes that make up a country's level of civil rights and political rights equally? Is academic freedom as important to democracy as an independent judiciary? Is it appropriate for Freedom House to weight the civil rights and political rights dimensions equally when coming up with a country's overall score? — isitappropriate for PolityIVto assume that moving from a 1toa 2ononeof its five attributes will have the same effect on a country's level of democracy as, say, moving from a 3 to a 4? Is Polity IV right to assume that all of its attributes measure democracy equally well? o measurement level + once having determined the best way to aggregate the attributes of a particular concept, we have to decide upon the most appropriate measurement level. There are three measurement levels: nominal + it classifies observations into discrete categories that must be mutually exclusive (not possible to assign any single case into more than one category) and collectively 30 exhaustive (the categories must be set up sothat all cases can be assigned to some category). A nominal measure is just a different way of naming cases. The DD measure is a nominal measure: it classifies / names countries as either democracies or dictatorships, given that it conceptualize democracy as a dichotomy; ii. ordinal + it allows us to rank / order cases along some dimension, to know whether a case has more or less of the thing we are measuring. For example, to say whether a country is more or less democratic than another country; iii. interval + it allows us to determine how much more or less a case has of the thing we are measuring. For example, to say how much a country is more or less democraticthan another country. Obviously, it requires that there be a standard unit of measurement. Both Polity IV and Freedom House are considered to be interval measures, given they conceptualize democracy as a continuum. e reliability > a reliable measure is one that repeatedly and consistently produces the same score for a given case when we apply the same measurement process / several people, when given the same rules for measuring democracy, all produce the same democracy score for a given country. What is the relationship between reliability and validity? A reliable measure is not necessarily a valid measure. A measurement process might produce results that do not change when repeated (reliable), but it does not necessarily mean that the measure is an accurate reflection of the concept under consideration (valid): for example, an invalid measure might be reliable because it repeatedly and consistently produces the same poor score for a given case. Similarly, a measure may be valid because on average it capturesthe underlying concept, but it may be unreliable because there might be a difference in any two attempts to measure the phenomenon. Obviously, it is preferred to have our measures to be both reliable and valid. n Na Reliable, but Not Valid Valid, but Not Reliable Valid and Reliable The reliability of a measure is likely to vary with the extent to which the measure depends on observable facts or subjective judgements: o the DD measure of regime type is likelyto be highly reliable because it is entirely based on observable facts. For example, one only has to be whether the chief executive and legislature are elected, whether there is more than one party, and whether there has been an alternation in power under identical electoral rules, to be able to code a country as a democracy or a dictatorship; Freedom House and Polity IV are likely to be less reliable because they are based on the subjective judgements of the individuals coding each country. For example, Freedom House asks country experts to code countries based on things such as fair electoral rules, equal campaigning opportunities, free and independent media, and reasonable self-determination. 31 e replicability > it refers to the ability of third-party scholars to reproduce the process through which a measure is created. Gary King has been one of the strongest proponents of replication: “[for] any empirical work enough information should be made available that a third party can comprehend, analyse, replicate, and evaluate the results without further information from the original author.” Replicability is important because it allows researchers who did not participate to the construction of a particular measure to independently evaluate its reliability and validity. At a minimum, replicability requires that scholars provide clear coding rules and make their disaggregated data available: o DD and Polity IV provide detailed and clear coding rules for constructing their measures of regime type; o Freedom House provides no coding rules for why a country might be given, say, a 30n one of its 25 questions instead of any other score. Disaggregated data refer to the scores given to each observation on the different attributes that are combined to produce a country's overall democracy or dictatorship score: o Polity IV makes all ofthe disaggregated data on each of its five attributes available to the public; o Freedom House has not historically provided the disaggregated data from its 25 questions, so there is no way to know what score a country was given on. This has led to conclude that the aggregate data offered by Freedom House has to be accepted largely on faith. In the last few years, a movement has emerged in political science calling for greater data access and research transparency (DA-RT). There are three components to DA-RT: o data access + it calls on scholars who have published research to make their data publicly available so long as doing so does not endanger human subjects, violate ethical norms and confidentiality agreements, or break the law; o productiontransparency > it callsonscholarsto be transparent about how their data were collected and how their measures were constructed; o analytic transparency + it calls scholars to be clear about exactly how their data and analyses support their evidence-based claims and inferences. 6. THE ECONOMIC DETERMINANTS OF DEMOCRACY AND DICTATORSHIP We will investigate how economic development and the structure of the economy influence the likelihood that a country will become and remain a democracy. 6.1 The classic modernization theory and the survival story The classic modernization theo! The modernization theory was originally developed by economists and economic historians and was later taken up by political scientists. Modernization scholars argue that all societies passthrough the same historical stages of economic development. In the aftermath of the post-World War Il decolonization, some countries were behind in this process of development. L, Attempts have been made to change the terminology used to describe these countries. They used to be called ‘primitive’, but scholars started to refer to them as ‘backward’, and then they became ‘third world” 32 Expected Probability of Regime Change Income Income Modernization Theory Survival Story — Transition to dictatorship ——-— Transition to democracy e both modernization theory and the survival story predict that the probability of a transition to dictatorship decreases as income increases (the solid lines in both panels slope down) + because high income helps the survival of democracy; e the modermnization theory predicts that the probability of a transition to democracy increases as income increases (the dotted line in the left panel slopes up) + because high income helps the emergence of democracy; e the survival story predictsthatthe probability of atransitionto democracy is unaffected by increasing income (the dotted line inthe right panel is flat) + because high income does not helpthe emergence of democracy. In sum, the positive relationship between high income and democracy is consistent with: e themodemizationtheory and its predictionthat countries are more likely to become and remain democratic as they develop and become richer; e the survival story and its prediction that countries are more likely to remain democratic as they develop and become richer. The modernization theory and the survival story are said to be competing theories. When political scientists have competing theories, they must try to deduce additional implications from their theories in the hope that these will help them decide which of the competing theories is the most consistent with the observed world. AIl four implications from modernization theory and survival story: Modernization theory and survival story 1. Democracy is more common in rich countries than poor countries. 2. Transitions to dictatorship become less likely as income increases. Modernization theory Survival story 3a. Transitions to democracy become more 3b. Transitions to democracy are unaffected likely as income increases. by increases in income. 4a. Regime transitions may or may not become 4b. Regime transitions become less likely as less likely as countries become richer. countries become richer. We now evaluate these implications using data from Przeworski and colleagues. Implication 1 > democracy is more common in rich countries than poor countries. Number of years all countries (country years) have lived under democracy or dictatorship at different levels of income between 1950 and 1990: 35 1,000 900 800 700 600 500 400 300 200 100 04 T T T T T T T T n 0 1 2 3: 4 5 6 7 8 9 GDP Per Capita (in Thousands of 1985 PPP US Dollars) Country Years =--- Years under dictatorship —— Years under democracy The graph shows that: ® when countries are very poor (say, when GDP per capita is below $2,000), almost 9 out of every 10 country years inthe data set are lived under dictatorship; ® when countries are relatively rich, however (say, when GDP per capita is above $8,000), virtually all the country years in the data set are lived under democracy; ® forthe countries in between (say, when GDP per capita is between $4,000 and $6,000), there are about as many country years under democracy as there are under dictatorship. Implication 2 > transitions to dictatorship become less likely as income increases + Implication 3 > transitions to democracy in particular: a) transitions to democracy become more likely as income increases, b) transitions to democracy are unaffected by increases in income. The probability of transitions to dictatorship, given a particular level of income, is calculated as follows: Pr (Transition to Dictatorship | Income Level) = Vumber of Transitions to Dictatorship Number of Transitions to Dictatorship Iricomie Level Number of Democratic Country Years The probability of transitions to democracy, given a particular level of income, is calculated as follows: Pr (Transition to Democracy | Income Level) = Number of Transitions to Democrac _—__ ad Income Level Number of Dictatorial Country Years Probability of transitions to democracy and dictatorship at different levels of income between 1950 and 1990: 36 s Ss È e È v E D ° H n 6 > 2 4 6 8 GDP Per Capita (in Thousands of 1985 PPP US Dollars) — Transitions to democracy Transitions to dictatorship a) the numbers are grey whenever a country is more likely to transition to dictatorship than democracy, b) the numbers are black whenever a country is more likely to transition to democracy than dictatorship. The graph shows that: e the probability of transitions to dictatorship (the grey dotted line) declines as income increases as predicted by both the survival story and modernization theory (implication 2). For example, transitions to dictatorship are 18 times more likely when GDP per capita is less than $2,000; e the probability of transitions to democracy (the straight black line) become more likely as income increases (implication 3a) asthe modernization theory predicts. For example, transitions to democracy is 6 times more likely when GDP per capita is greaterthan $6,000. Thus, this point falsifies implication 3b of the survival story. Implication 4 + regime transitions in general: a) regime transition may or may not become less likely as countries become richer, b) regime transitions become less likely as countries become richer. The probability of a regime transition, given a particular level of income, is calculated as follows: Pr (Regime Transition | Income Level) = Number of Transitions to Democracy or Dictatorship | — Income Level Number of Country Years This equation tells us that the probability of a regime transition, given a particular level of income, is equal to the number of transitions to democracy or dictatorship at that income level divided by the number of country years that could have transitioned at that income level. Probability of regime transitions at different levels of income between 1950 and 1990: 0.040 0.035 0.030 0.025 0.020 0.015 y of a Regime Transition 0.010 0.005 04 DI 2 4 6 8 GDP Per Capita (in Thousands of 1985 PPP US Dollars) — Transitions to democracy or dictatorship regime transitions in general 37 scholars to explainthe establishment of modern parliamentary democracy in Britain during the Glorious Revolution of 1688. Bates and Lien argue that the introduction of this more limited state occurred earlier in England than it did in France because ofthe unique structure ofthe economy that early modernization had produced in England. Why? We answer to this question with the EVL game. In the prehistory of the EVL game, the Crown, under the exigencies of war, has confiscated or taxed the assets of a segment of the elite represented by the Parliament / the Parliamentarians. The Parliamentarians, from their newfound position of strength, have three options: 1. exit / disinvest+ to disinvest in the economy: to take the remaining assets and do everything they can to shield them from further confiscation. If the Parliamentarians no longer invest, the economy is likely to stagnate, and there will be less forthe Crown to tax or confiscate inthe future; N voice / demand limits + to petition the Crown for protections against future confiscations / for limits on future predatory behaviour (say, by granting Parliament the right to veto all future increases in taxation or by establishing an independent judiciary to police the Crown's behaviour) in exchange for a promise to continue investing their assets in the economy. If the Parliamentarians choose to use voice, the Crown has two options: respond / accept limits (on its power to tax) > the Parliamentarians will happily continue to invest and the economy will grow; ii. ignore / reject limits. If the Crown chooses to ignore, the Parliamentarians have two options: a) loyalty /continue investing as before + the economy will continue to grow; b) exit / withdraw substantial portions of their assets from the market + the economy will stagnate. 3. loyalty /invest > to continue investing as they did before the confiscation and pay taxes. Parliamentarians Exit: Disinvest Loyalty: Invest 01: Unlimited government, Voice: Demand limits 02: Unlimited government, stagnant economy | growing economy Crown Respond: Accept limits | Ignore: Reject limits 03: Limited government, growing economy Parliamentarians Disinvest Invest (04: Unlimited government, 5: Unlimited government, stagnant economy growing economy Scenario 1 > the Parliamentarians have a credible exit threat (E >0) and the Crown is dependent (L> 1). 40 Parliamentarians Itisassumedthatc>0; E<1-c E<0;andL>1. Exit: Disinvest Loyalty: Invest The subgame perfect equilibrium is (demand 61 Voice: Demand limits 0141 limits, disinvest; accept limits). Crown The observed outcome is a limited government with a growing economy. Respond: Accept limits Ignore: Reject limits 1-GL T Parliamentarians Disinvest Invest the observed outcome E-G1 0-c1+L The Crown decides to accept limits on its predatory behaviour because it knows that it is dependent on the Parliamentarians for its money and because it knows that the Parliamentarians will disinvest and exit if it rejects the limits. Knowingthat their petition will be effective, the Parliamentarians use voice and demand limitsfromthe Crown. This particular scenario helps to explain why the English monarchy in early modern Europe accepted limits on its predatory behaviour because it depended on elites with credible exit threats (mobile assets). Scenario 2 > the Parliamentarians do not have a credible exit option (E < 0: immobile assets) and the Crown is dependent (L>1). Parliamentarians Itis assumedthatc>0;E<1-c;E>O;andL>1. Exit: Disinvest Loyalty: Invest ] The subgame perfect equilibrium is (invest, invest; E1 Voice: Demand lirpits 0,141 reject limits). The observed outcome is an unlimited government Crown with a growing economy. Respond: Accept limits Ignore: Reject limits 1-gL Parliamentarians Disinvest Invest E-g1 0-G14+L The Crown will reject any demands to limit its predatory behaviour because it knows that, although it is dependent on the Parliamentarians for money, the Parliamentarians will continue to invest and pay their taxes even in a predatory environment due to the fact that they do not have a credible exit option. Knowing that the Crown will ignore their petitions, the Parliamentarians simply continue to invest and pay their taxes since the beginning of the game. This scenario helps to explain why the French monarchy in early modern Europe did not accept limits on its predatory behaviour because it depended on elites who did not have credible exit threats (immobile assets). Scenario 3 > the Parliamentarians have a credible exit option (E > 0) and the Crown is autonomous (L < 1). 41 e The observed outcome is an unlimited government with a stagnant economy. The Parliamentarians will choose to exit and disinvest inthe economy: they have a credible exit option and they know that the Crown does not depend on them. Scenario 4 > the Parliamentarians do not have a credible exit option (E < 0) and the Crown is autonomous (L< 1). e The observed outcome is an unlimited government with a growing economy. The Parliamentarians will choose to continue investing and paying theirtaxes: they do not have a credible exit option and they know that the Crown does not depend on them. Summary of outcomes in the EVL game: Crown ls autonomous Is dependent L<1 LSA Parliamentarians Have a credible exit threat Poor dictatorship Rich democracy (mobile assets) (unlimited government, (limited government, E>0 stagnant economy) growing economy) Have no credible exit threat Rich dictatorship Rich dictatorship (fixed assets) (unlimited government, (unlimited government, E<0 growing economy) growing economy) In conclusion, representative government (of which limited government / democracy is an example) is more likely to emerge and survive when: a) the state depends on a segment of society (economic elites / Parliamentarians), and b) this segment of society (economic elites / Parliamentarians) has a credible exit option (mobile assets). Hobbes saw the creation of a state as a solution to the state of nature. Theorists such as Locke recognized that the creation of the state might solve the problem that citizens have with each other, but by surrendering control over the means of violence to the state, what was to prevent the state from using its power against its citizens? The argument we have just presented illustrates that there are some conditions under which the state will voluntarily agree to accept limits on its predatory behaviour; under these conditions, no one will need to guard the guardian because the guardian will guard itself. Natural resources and democracy This variant of modernization theory also provides an explanation for the political resource curse. According to the political resource curse, countries that depend on revenue from natural resources, such as oil, diamonds, and minerals, will find it difficultto become democracies. You might think that having natural resources would be a blessing as these resources provide access to ‘free’ or ‘unearned’ income that can be used to improve the material well-being of citizens and build democracy. The empirical evidence, though, consistently shows that countries with natural resources are more likely to be dictatorships than democracies and that they are also more prone to corruption, poor governance, and civil war. Far from being a blessing, natural resources appear to be a curse. Estimated value of oil and gas produced per capita in 2009 in current dollars: aid pessimists, who think that foreign aid has a negative effect on the democratization process. Our EVL game showed that democracy is unlikelyto emerge when governments are autonomous from their citizens. Foreign aid hinders the emergence of democracy because it increases the autonomy of recipient governments from the demands of their citizens by freeing governments from the need to raise taxes and providing them with access to ‘slack resources’ that can be strategically used to reward supporters and co-opt opposition groups. One may see foreign aid as a way to ease the pain of those citizens living under a dictatorial rule. However, many studies show that foreign aid to dictatorships harms the welfare of the average citizen and helps dictators stay in office through corruption and exploitation ratherthan through the production of effective public policy. This reminds of the reasoning behind the political resource curse. Many scholars claim that income from foreign aid is essentially the same as income from natural resources because both represent forms of unearned income that increase the autonomy of governments relative to their citizens. In fact, some people speak of a foreign aid curse, similar to a resource curse. This line of reasoning, though, overlooks some significant differences between natural resources and foreign aid: È natural resource revenue makes governments more autonomous in a general sense; foreign aid makes governments less dependent on their citizens, but it makesthem more dependent on their foreign aid donors + aid donors will be able to exert influence on these dependent recipient countries whenever they decide to credibly threaten to withdraw their aid. What exactly do donor countries want when they give aid? Donor countries often want to achieve strategic goals with their foreign aid. For example, donor countries might use foreign aid to encourage recipient countries to attack terrorist groups, such as ISIL and the Taliban, to prop up ‘friendly’ regimes in hostile regions of the world, orto obtain privileged access to economic markets for their domestic companies. However, many donor countries do demand democratic and other political reforms as a condition for continued foreign aid. Empirical evidence, though, suggests that these types of conditions are rarely enforced by the donor country. One reason is that donor countries, while publicly stating that they desire democratization, often have little actual interest in recipient countries becoming democratic because it is typically cheaper to get a dictatorship rather than a democracy to adopt ‘friendly’ policies. This has led some scholars to claim that foreign aid is effectively ‘unconditional’ in practice. If this is true, then we should believe that aid donors have only interest in achieving their strategic goals, not in promoting democratization. When can donor countries credibly threaten to withdraw their aid? When the aid donor has a strategic interest inthe recipient country, threatsto withhold aid are not credible. lt ishard to imagine that the United States would withdraw significant amounts of aid from countries such as Iraq and Afghanistan iftheyfailto democratize because the governments in these countries are strategically importanttothe United States in terms offighting international terrorism and achieving stability inthe Middle East. Recipient countries that are strategically important are often aware of this fact and feel that they can safely ignore any demands for political reform. When the donor country has no strategic interests in the recipient country, threats to withhold aid are credible. In sum, foreign aid can promote democratization only under all these extremely demanding conditions: the recipient country is dependent on the foreign aid; the aid donor wants to promote democratic reform in the recipient country; 45 e the aid donor can credibly threaten to withdraw aid if its demands for reform are not met. However, if political reforms do actually occur in the recipient country, they are likely to be limited in scope. Most dictatorial leaders in recipient countries will try to avoid introducing meaningful reforms that put their authority in jeopardy. Instead, they will try to implement superficial reforms that are sufficient to satisfy their foreign aid donors but do not actually threaten their hold on power. Dietrich and Wright present evidence consistent with this expectation. They find that foreign aid tends to increase de jure (by law, on paper) political competition inthe recipient country but not de facto (in fact, in practice) political competition. For example, they find that while foreign aid might encourage the emergence of multiparty politics, the resulting multiparty politics are characterized by weak opposition parties and the incumbent leaders continue to dominate. Inequality and democracy Tocqueville argued that economic equality was important for the introduction and persistence of democratic institutions. Economic inequality undermines democracy (the claim) because it produces political competition between the rich and the poor (the basic argument) as it provides incentives forthe poorto redistribute wealth from the rich. The ability of the poor to take wealth away from the rich simply by voting makes democracy appealing to the poor but costly to the rich. As a result, the economic elites will step in to block attempts at democratization or to conduct coups to reverse democratization in highly unequal societies. However, the argument lacks strong empirical support. Some studies find a negative relationship between inequality and democracy, but most find no relationship at all. But they also fail to find a consistent negative relationship between inequality and democracy. Economic elites who have credible exit options (mobile assets) can force a dependent state to accept limits on its predatory behaviour. That these economic elites have credible exit options and can realistically withdraw their much- needed investment inthe economy also helps to explain why the poor will not vote to expropriate them. Therefore, economic inequality should be considered bad for democracy only in countries where the economic elites have immobile assets and therefore lack credible exit threats, not in countries where the economic elites have mobile assets and are willing to keep democracy safe. Evidence in support of this line of reasoning comes from Freeman and Quinn, who find that unequal dictatorships that are financially integrated into the global economy are much more likely to experience democratic transitions than unequal dictatorships that are financially closed. There is also a difference between income inequality (which promotes democratization) and land inequality (which hinders democratization): e income inequality generally increases along with the creation of a large middle class, a group that tends to have mobile assets; e landinequality usually goes hand in hand with a large landed aristocracy, a groupthat tendsto have immobile assets and that, for this reason, fears democratization. Economic performance In addition to providing predictions about whether countries are likely to be democratic or dictatorial, our EVL game also offers predictions about whether they will have growing or stagnant economies. The outcomes indicate that the economic performance of democracies will be good, whereas that of dictatorships will be much more varied. 46 6.3 Some empirical evidence First, we briefly describe how to interpret key pieces of information typically found in a table of statistical results: e the dependent variable (Y), which is listed at the top of the table, is the thing we want to explain. It is the effect in a proposed cause-and-effect relationship; e the independent variables (X), which are listed in the first column, are the things we think will explain the dependent variable. They are the causes in a proposed cause-and-effect relationship; e next to each independent variable, in the other columns, is a coefficient. The sign of the coefficient tells us the slope / pattern of the relationship between X and Y: o a positive coefficient indicates that there is a positive relationship between X and Y; o a negative coefficient indicates that there is a negative relationship between X and Y; o the coefficient will be zero if there is no relationship between X and Y. e arethe patterns described by the coefficients likely to be found only in this particular data set or outside of this data set / generally too? Each coefficient comes with a corresponding standard error beneath it in parentheses. The standard errors are measures of uncertainty, and they help usto determine how confident we should be in our results. We tend to be confident that we have found a pattern in the data that is likely to be found more generally when the coefficient is bigger than twice the size of the standard error; e starsare placed next to the coefficients, where the size of the coefficient is sufficiently large relative to the size of its standard error: these independent variables are statistically significant. Otherwise, the independent variables are statistically insignificant; e the constantis also called ‘intercept’ because it isthe point where the line intercepts the vertical Y-axis, and it indicates the value of Y when X is 0; e observationsare units of analysisthat refer tothe entitiesthat we are talking about in ourtheory: individuals, groups, countries, country-years (number of countries multiplied by number of years). Statistical analysis is a lot more useful when the number of observations is high; e loglikelihood isone of the measures used to say how good a model explains the data. Has the IV coefficiet star(s) **? Do nothing Interpret the sign +/- Coefficient is not (significantly) Coefficient is significantly different different from zero (random) from zero IV is not correlated with DV IV is significantly correlated with DV, Let us evaluate some of the arguments that have been presented in this chapter using statistical analyses. Argument 1. How a country's status as an oil producer, its income, and its economic growth affect the probability that it will become a democracy. Economic determinants of democratic emergence (Przeworski and colleagues: all countries from 1946 to 1990): 47 e Howmuchless likely would it be for Burkina Faso to remain a democracy in 1988 if its growth rate had been the same as that of the United States (2.55%) instead of -2.15%? The answer is 4% more likely to remain a democracy if its growth rate per capita had increased from -2.15% to 2.55%. ‘7. THE CULTURAL DETERMINANTS OF DEMOCRACY AND DICTATORSHIP We will investigate how culture influences the likelihood that a country will become and remain a democracy. 7.1 Classical cultural arguments: Mill and Montesquieu Cultural arguments regarding democracy typically fall into two categories: ® primordialist arguments treat culture as something that is objective and inherited. It is culture that affects political behaviour by providing ideological guidelines for collective action, rather than political behaviour that affects culture. Culture is something that has been fixed since primordial times. As a result, political institutions may not be compatible with all cultures (democracy may not be compatible with all cultures); e constructivist arguments treat culture as something that is constructed or invented. Like primordialists, constructivists agree that it is culture that affects political behaviour, rather than political behaviour that affects culture. However, they recognize that cultures are malleable and are not given once and for all, that cultures can change in response to social, economic, and political actors, and that the speed with which cultures can change is likely to vary from culture to culture. As a result, some cultures will find it easier to adopt a political institution than others (some cultures will find it easier to adopt democracy than others). The notion that different cultures were suited to different political institutions, such as democracy and dictatorship, has a long history: ® in472B.C,, Aeschylus contrasted the authoritarianism associated with the people of Asiawith the democracy found in Athenian Greece in his play The Persians. His views would later be echoed inthe Asian values debate inthe 19905; e inthe 18° century, Montesquieu wrote that: o a)monarchy was most suited to European states, b) despotism was most suited to the Orient, and c) democracy was most suited to the ancient world; o the best government for a given country was one which “leads men by following their propensities and inclinations” and which “best agrees with the humour and disposition of the people in whose favour it is established”; o political institutions “should be in relation tothe climate of each country, to the quality of its soil, to its situation and extent, to the principal occupation ofthe natives, whether husbandmen, huntsmen, or shepherds: they should have relation to the degree of liberty which the constitution will bear; to the religion of the inhabitants, to their inclinations, riches, numbers, commerce, manners, and customs”; o it can be only by chance that the political institutions of one country can successfully be exported to another. e inhisdiscussion To what extent forms of government are a matter of choice, John Stuart Mill statedthat “no one believes that every people is capable of working every sort of institution. [...] Nothing but foreign force would induce a tribe of North American Indians to submit to the restraints of a regular and civilized government.” Even those people who recognized the benefits of a civilized government might still have to 50 live under authoritarianism if they did not have the required characteristics to support a better system of government, which are: o “mora public authorities in the repression of evil-doers”; or “mental habits”, such as the willingness to “co-operate actively with the law and the o a certain degree of development characterized, for example, by a press capable of propagating public opinion and a tax system “sufficient for keeping up the force necessary to compel obedience throughout a large territory”. Therefore, legislators should take account of “pre-existing habits and feelings” when creating political institutions in a country. “People are more easily induced to do, and do more easily, what they are already used to”, but “people [...] learn to do things new to them” because culture is inherently malleable and, as a result, people can learn to live with democracy. As you can see, Mill asserted a constructivist cultural argument: he did not think that certain cultures are incompatible with political institutions: he was highly critical of those who believed that culture prevented political actors from choosing the institutions they desired. The cultural arguments put forth by both Montesquieu and Mill were later incorporated into the cultural modernization theory: modernization theory predicts that “immature” societies will eventually become “mature” societies as they develop economically; cultural modernizationtheory predicts that societies with primitive cultures will eventually become societies with civilized cultures as they develop socioeconomically, and onlywhen this happens societies are ready for democracy. The claims made by Montesquieu and Mill regarding culture and democracy illustrate several potential problems that characterize some cultural arguments to this day: what is it about culture that matters for democracy? Mill and Montesquieu provided a whole host of cultural things (religion, customs, morals, manners, marital institutions, etc.) and non-cultural things (the climate, the quality of the soil, and the economy) that might affect the emergence and survival of democracy. But most of these things they point out are left so vague or nonspecific that they become non-falsifiable, and hence, non-scientific. What particular morals are incompatible with democracy? Which customs are problematic? \f culturalist arguments are to have any explanatory power, they must distinguish and specify what it isthat matters; what is the causal relationship between cultural, economic, and political factors? Mill and Montesquieu clearly believed that economic development and culture both mattered for democracy. From what they wrote, however, it is hard to discern what they thought the exact causal relationship was between these factors. Which way does the causal arrow go? Does culture cause political institutions, such as democracy, to emerge and survive? Does it also cause economic development? Or do political institutions and economic development cause culture? If culture is a cause, does it cause the emergence of democracy, or does it affect only the survival of democracy? If culture does cause democracy, is it a necessary or a sufficient condition? Some potential causal relationships between culture, economic development, and democracy: 51 & Democracy b. Democracy IT Culture Culture Economic Economic development development Culture s»— ECONOMIC development. -s— DEMO Cr ACY Economic development sessi Cult Ure si» DE MOCFACY Economic development. Democracy Culture Economic development ra» DEMMOC1 2Cy e CUTE Economic development === DeMOcracy Culture Over the years, there has been considerable debate about this: o scholars who argue that economic development produces cultural change and that it is cultural change that produces democracy + values story (Inglehart and Welzel): Economic development rami CU lt Ure mmm» De MOC!rACY o scholarswho argue that economic development produces democracy and that it is experience with democracy that produces cultural change + institutional story (Barry, Muller, Seligson): Economic development. sno» DEMOCrACy mio» CUltUrE 7.2 Does democracy require a civic culture? Gabriel Almond and Sidney Verba Almond and Verba recognized three basic types of political culture in the world: e parochial culture > compatible with the traditional political systems: African tribes; e subject culture + compatible with centralized authoritarian institutions: Soviet-dominated Eastern Europe; e participant / civic culture + compatible with democracy. Broadly speaking, a civic culture refers to a shared cluster of attitudes that are thought to promote democracy and democratic performance (only a “civic culture” can provide the “psychological basis of democratization”). However, there has never been complete agreement on the precise cluster of attitudes thought to compose a civic culture. For Almond and Verba, political culture is something that captures how individuals think and feel about the political system. Individuals’ thoughts and feelings depend on whether a) individuals believe they can influence political Surveys and comparative research Surveys are increasingly being used to examine many important questions in comparative politics. The most commonly used survey today is the World Values Survey (which provided the data forthe cultural map of the world shown above): it conducted interviews in 80 societies and it addresses issues of sociocultural and political change. Example of question from the World Values Survey: “Democracy may have problems, but it is better than any other form of government. Could you please tell me if you a) strongly agree, b) agree, c) neither agree nor disagree, d) disagree, or e) strongly disagree ?”. Many people believe that mass support for a particular system of government provides political systems with the legitimacy they need to operate effectively: for example, mass support for democracy is seen by some as essential in delegitimizing dictatorial rule and legitimizing democratic rule. Thus, a low level of public support reported in questions like this is seen as a sign of democratic instability or collapse. Researchers who use surveys often confront certain problems, two of which arise when: surveys address sensitive topics + in survey research we assume that the respondents are answering the survey questions truthfully, but unfortunately, respondents often lie or hide their true beliefs. This problem is commonly referred to as ‘social desirability bias': the tendency of survey respondents to answer questions in a manner that will be viewed favourably by others / to overreport ‘good behaviour” and underreport ‘bad behaviour. For example, it is hard to use surveys to study the emergence of democracy because surveys would need to be conducted in dictatorships. A respondent in a dictatorship may not feel comfortable revealing her true preferences when asked about the value of democracy because democracy is a sensitive topic in most authoritarian regimes, so, he may underreport its enthusiasm for democracy for fear to be punished for expressing its sincere beliefs. Indeed, these concerns can even lead to certain types of people not participating in surveys in the first place, with the result that these surveys are no longer representative; individuals in different countries often understand survey items in different ways or evaluate survey items using different scales + this problem is commonly referred to as ‘differential item functioning’ (DIF) and it applies in particular to cross-national surveys, such as the World Values Survey. This problem is especially relevant for survey research related to democracy because democracy means different things to different people around the world: democracy may mean economic and political equality for some, but it may simply mean holding competitive elections for others. These different views of democracy should not come as a surprise, given that we have already seen that even political scientists disagree about whether to employ a minimalist or substantive view of democracy in their own work. If experts cannot agree on what they mean by democracy, why would we expect individuals in different countries to have the same concept in mind when answering survey questions about it? Political scientists have begun to develop methods to get around these problems: Blair describes four basic methods for addressing sensitive topics with surveys: 1. the first method involves the use of practices that build trust with respondents. Some scholars recommend using interviewers who share similar demographic characteristics, such as age, gender, or ethnicity, with the respondent. Adida and colleagues find that respondents in Africa give systematically different answers to survey questions ifthe people who interview them are co-ethnics as opposed to non-co-ethnics. 55 N Others have highlighted the importance of allowing respondents to report their answers in private rather than aloud to an interviewer. In a study of caste-sensitive attitudes in India, Chauchard provided his respondents with MP3 players so that they could self-report their answers in private; the second method involves the use of randomized response techniques or devices (such as a coin or a die) to guarantee the confidentiality of the respondents’ answers. For example, before answering a sensitive survey question that requires a ‘yes’ or ‘no’ answer, a respondent might be asked to rolla die in private. Ifthe die shows a 1, the respondent is told to report ‘ves’. If the die shows a 6, the respondent is told to report ‘no’. If the die shows a 2, 3, 4, or 5, then the respondent is told to answer the question truthfully, either ‘yes’ or ‘no’. As you can see, the interviewer never knows whether a given individual response is true or not; hence, the confidentiality of the individual response is guaranteed. The interviewer knows the probability with which truthful answers are given (ifthe respondents follow the instructions) and can therefore calculate the overall numbers of respondents who said ‘yes’ and ‘no’ to the sensitive question. Randomized response techniques have been used to study a variety of phenomena, such as corruption within bureaucracies, cheating by undergraduates, sexual attitudes, and the prevalence of xenophobia and anti-Semitism; the third method involves the use of list experiments to protect the confidentiality of individual responses by mixing sensitive items into lists that include non-sensitive items. We divide the respondents into two groups: o some survey respondents are presented with a list of non-sensitive control items and are then asked to indicate the number of items with which they agree; o some other survey respondents are presented with the same list of non-sensitive control items except that a sensitive item is now also included and are then asked to indicate the number of items with which they agree. By comparing the number of items that respondents agree with across these two randomly selected groups, it is possible to identify the level of support forthe sensitive item. This method does not always guarantee that individual responses are kept confidential. This is because a respondent who answers that she agrees with none or all of the items inthe list containing the sensitive item reveals her preferences about the sensitive item. As a result, researchers need to be careful about what items they include in the list of control items. List experiments have been used to study a variety of phenomena such as public support for coalition forces in Afghanistan, vote buying in Nicaragua, and the impact of citizen preferences on policy in China; the fourth method involves the use of endorsement experiments to protect the confidentiality of individual responses by mixing attitudes toward a sensitive political actor with attitudes toward one or more policies: © in an endorsement experiment, some survey respondents are asked to give their attitudes toward a policy that a sensitive political actor has endorsed. The responses to this question mix the preferences of the respondents toward the policy and their preferences toward the sensitive actor; 56 o other survey respondents are asked to give their attitudes toward the same policy but without any mention of the endorsement. The responses to this question isolate the preferences of the respondents toward the policy. By comparing the responses across the two groups, it is possible to identify just the preferences toward the sensitive political actor. Endorsement experiments have been used to study a variety of phenomena such as attitudestoward the Taliban in Afghanistan and militants in Pakistan. ® political scientists attempt to deal with the problem that respondents do not always comprehend survey questions in the same way through the use of anchoring vignettes. Anchoring vignettes are (usually brief) descriptions of hypothetical people or situations useful for survey questions that use an ordinal scale (say, strongly disagree, disagree, neutral, agree, strongly agree) such as the democracy question seen earlier. Anchoring vignettes have been used by the World Health Organization to examine various health indicators, the World Bank has used them to investigate economic welfare, and other scholars have used them to look at things like corruption, political efficacy, and women's autonomy. The development of these new techniques for overcoming the challenges of using survey responses to measure cultural differences is promising, but their application to the study of political culture has just begun, and only time will tell whether they will work. 7.3 Religion and democracy In The clash of civilizations and the remaking of world order, Samuel Huntington wrote that “the fundamental source of conflict in this new world will not be primarily ideological or primarily economic. The great divisions among humankind and the dominating source of conflict will be cultural. [...] the principal conflicts of global politics will occur between nations and groups of different civilizations. The clash of civilizations will dominate global politics.” He defines ‘civilization’ asthe “highest cultural grouping of people and the broadest level of cultural identity people have short of that which distinguishes humans from other species”: Western Christian, Confucian, Japanese, Islamic, Hindu, Slavic-Orthodox, Latin American, African, and others. The exact number is ambiguous, but on the whole, civilizations seem to be coded primarily in regard to religion, although linguistic differences and geographic proximity seem to play a role in some cases. He argues that “Western concepts differ fundamentally from those prevalent in other civilizations. Western ideas of individualism, liberalism, constitutionalism, human rights, equality, liberty, the rule of law, democracy, free markets, the separation of church and state, often have little resonance in Islamic, Confucian, Japanese, Hindu, Buddhist or Orthodox cultures.” The widespread Western belief in the universality of the West's values and its insistence on imposing these values through democratization efforts will only antagonize other civilizations and lead to conflict. These conflicts will be less amenable to diplomacy and peaceful resolution than previous economic and ideological conflicts because cultural differences are less mutable and less easy to compromise. In effect, Huntington asserted a primordialist argument: certain cultures are incompatible with democracy. Are some religions incompatible with democracy? Debates concerning the relationship between religion and democracy have a long history: e Protestantism has historically been seen as a religion that encourages democracy: o Max Weber is commonly thought to have provided the first argument linking Protestantism with democracy in The Protestant ethic and the spirit of capitalism: Protestantism encourages economic 57 e 140fthe 47 countries with a Catholic majority were coded as Free, whereas 16 were coded as Not Free; e llofthe 16 countries with a Protestant majority were coded as Free, whereas only one was coded as Not Free. This would suggest that Islam and Catholicism are bad for democracy and Protestantism is good for democracy. However, in 2004, Catholic countries had higher Freedom house scores than Catholic countries in 1976: e 400fthe 57 countries with a Catholic majority were coded as Free, whereas only 3 were coded as Not Free. It turns out that most of the arguments claiming that particular religions or civilizations are incompatible with democracy are implicitly based on observations of the world at a particular point in time. Theory construction came after observing the world: e mostofthe arguments that Protestantism is good for democracy occurred when most of the democracies inthe world were Protestant; e most of the arguments that Catholicism is bad for democracy occurred when there were few Catholic democracies in the world; ® today, people argue that Islam is bad for democracy because we do not have many Islamic democracies. Are some religions incompatible with democracy? A new test We know that today Protestant and Catholic countriestend to be democratic and that Muslim countries tend to be authoritarian. This has already been demonstrated by some of the studies we just mentioned. However, this does not establish a clear link between these religions and the prospects for democracy. We also know that democracy originated in Protestant countries. But our question is not about where democracy originated. What we really want to know is whether democracy is more or less likely to emerge and survive in countries that are dominated by Protestants, Muslims, or Catholics. This requires examining the effect of religion on democracy across time. To do this, we need to know what effect being a Protestant, Catholic, or Muslim country has on a) the probability of becoming democratic, and b) the probability of remaining democratic. In other words, we would like to test the following hypotheses: e Protestant hypothesis + countries with a majority Protestant population are more likely to become and stay democratic; e Catholic hypothesis + countries with a majority Catholic population are less likely to become and stay democratic; e Islamic hypothesis + countries with a majority Muslim population are less likely to become and stay democratic. 60 Religion of majority Muslim Protestant Catholic Countries Afghanistan, Albania, Algeria, Azerbaijan, Bahrain, Bangladesh, Brunei, Comoros, Djibouti, Egypt, Eritrea, Gambia, Guinea, Indonesia, Iran, Iraq, Jordan, Kuwait, Kyrgyzstan, Lebanon, Libya, Malaysia, Maldive Islands, Mali, Mauritania, Morocco, Niger, Oman, Pakistan, Qatar, Saudi Arabia, Senegal, Somalia, Sudan, Syria, Tajikistan, Tunisia, Turkey, Turkmenistan, United Arab Emirates, Uzbekistan, (Republic of) Yemen Angola, Antigua, Bahamas, Barbados, Denmark, Fiji, Finland, Iceland, Liberia, Marshall Islands, Namibia, Norway, Papua New Guinea, St. Kitts and Nevis, St. Vincent, Solomon Islands, South Africa, Sweden, Tonga, United Kingdom, United States, Vanuatu, Western Samoa Andorra, Argentina, Armenia, Austria, Belgium, Belize, Bolivia, Brazil, Burundi, Cape Verde, Chile, Colombia, Congo, Costa Rica, Croatia, Cyprus, Dominica, Dominican Republic, Ecuador, El Salvador, Equatorial Guinea, France, Gabon, Greece, Grenada, Guatemala, Haiti, Honduras, Hungary, Ireland, Italy, Kiribati, Liechtenstein, Lithuania, Luxembourg, Macedonia, Malta, Mexico, Micronesia, Nicaragua, Panama, Paraguay, Peru, Philippines, Poland, Portugal, Romania, Rwanda, St. Lucia, San Marino, Sao Tomé and Principe, Seychelles, Slovakia, Slovenia, Spain, Uruguay, Venezuela, Yugoslavia Some scholars have argued that cultural diversity is bad for democracy because democracy can be sustained only if citizens can coordinate their beliefs about when the government has transgressed and when they should do something about this transgression. In many ways, this coordination of beliefs might be considered a ‘democratic culture’, something that is necessary for democracy to emerge and survive. It seems reasonable to think that this type of coordination is likely to be more difficult when there are many cultural groups in society. Other scholars have argued that ethnic diversity is bad for democracy because it makes reaching compromises difficult and because it raises the risk of intercommunal violence. As a result, we also evaluate the following hypotheses in our empirical analyses: e ethnic group hypothesis + countries with a large number of ethnic groups are less likely to become and stay democratic; e religious group hypothesis + countries with a large number of religious groups are less likely to become and stay democratic; e cultural group hypothesis + countries with a large number of cultural groups are less likely to become and stay democratic. 61 Effect of a Muslim, Protestant, or Catholic majority on the probability that a country will become democratic between 1950 and 2000: Dependent variable: Probability of being a democracy this year if country was a dictatorship the previous year Independent variables Model 1 Model 2 Model 3 Model 4 Model 5 Muslim majority -0.28#* -0.18 -0.23 -0.25 0.18 « (0.12) (0.16) (0.17) (0.19) (0.16) «— Protestant majority -0.56 -0.42 -0.40 -0.45 -0.43 (0.35) (0.38) (0.38) (0.39) (0.38) Catholic majority 0.334** 031*** 0.26** 0.26** 031** (0.10) (0.12) (0.12) (0.13) (0.13) GDP per capita 0.00004* 0.00003* 0.00003* 0.00004* (0.00002) (0.00002) (0.00002) (0.00002) Growth in GDP per capita -0.02** -0.02%* -0.02** -0.02** (0.01) (0.01) (0.01) (0.01) Oil production -0.15 -0.12 -0.13 -0.15 (0.18) 0.19) (0.19) (0.18) Effective number of ethnic groups -0.02 (0.02) Effective number of religious groups -0.06 (0.09) Effective number of cultural groups 0.02 (0.08) Constant -2.06*** -2.05%%% 1.940 2191608 —2.06%*% (0.07) (0.10) (0.13) (0.23) (0.19) Number of observations 4,379 2,578 2,563 2.578 2,563 Log-likelihood —418.75 -318.64 -317.85 -318.46 -318.35 Model 1 examines how having a Muslim, Protestant, or Catholic majority affects the emergence of democracy without taking anything else into account: e the coefficient on ‘Muslim majority is negative and statistically significant, meaning that having a Muslim majority reduces the probability that a country becomes a democracy; e the coefficient on ‘Protestant majority' is statistically insignificant, meaning that we cannot confidently rule out the possibilitythat there is no relationship between Protestantism and the emergence of democracy; e the coefficient on ‘Catholic majority' is positive and statistically significant, meaning that having a Catholic majority increases the probability that a country becomes a democracy. If we looked only at the results from Model 1, we would have to conclude that having a Muslim majority is bad for the emergence of democracy. We know, however, thatthese countriestend to be poorerthan most other countries. We also know that poor countries are less likelyto become democratic than rich countries. Thus, it might be the case that Muslim countries are less likely to become democratic not because they are Muslim but because they are poor. To test this possibility, we include in Model 2 the three economic variables used to examine the economic determinants of democracy: GDP per capita, economic growth, and oil production: e the coefficient on Muslim majority' is statistically insignificant, meaning that we cannot confidently rule out the possibility that there is no relationship between Islam and the emergence of democracy. The evidence suggests that Muslim countries are less likely to become democratic not because they are Muslim but because they are poor. If these countries can develop economically and become wealthier, then, there is no reason to thinkthat these countries will not become democratic. Model 3, 4 and 5: e the coefficients on our ethnic, religious, and cultural diversity variables are statistically insignificant, meaning that we cannot confidently rule out the possibility that there is no relationship between these forms of diversity and the emergence of democracy. 62 e the modal offer /the most common offer in student populations playing the Ultimatum Game is normally almost 50%, with the mean / average offer varying somewhere between 40% and 45%. Offers of less than 20% of the pie are rejected by responders about half the time; e themodalofferinstudent populations playing the Dictator Game is normally 0%, with the mean offer varying somewhere between 20% and 30%: ® asubstantial portion of the students do not approximate the theoretical Homo economicus. Do some cultures exhibit behaviour that more closely resembles the theoretical predictions than others? Although there are cultural differences among students in different countries, these differences are quite small compared with the range of cultural environments that exist inthe world. A group composed primarily of anthropologists and economists decided to conduct experiments using the Ultimatum and Dictator Games in 15 small-scale societies in 12 countries on 5 continents. These societies exhibited a wide range of cultural and economic environments: 3 foraging societies, 6 slash-and-burn horticulture societies, 4 nomadic herding groups and 2 sedentary small-scale agriculture societies. Group Machiguenga Quichua Achuar Hadza Aché Tsimané AU Gnau Mapuche Torguud Khazax Sangu (farm/herd) Orma Lamelara Shona Country Peru Ecuador Ecuador Tanzania Paraguay Bolivia Papua New Guinea Papua New Guinea Chile Mongolia Mongolia Tanzania Kenya Indonesia Zimbabwe Environment Tropical forest Tropical forest Tropical forest Savanna-woodlands Semi-tropical woodlands Tropical forest Mountainous tropical forest Mountainous tropical forest Temperate plains High-altitude desert, seasonally flooded grassland High-altitude desert, seasonally flooded grassland Savanna-woodlands, seasonally flooded grassland Savanna-woodlands Tropical island coast Savanna-woodlands Economic base Horticulture Horticulture Horticulture Foraging Horticulture and foraging Horticulture Foraging and horticulture Foraging and horticulture Small-scale farming Pastoralism Pastoralism Agro-pastoralists Pastoralism Foraging-trade Farming 65 Offers made in the Ultimatum Game: the size of the bubble Lamelara +» at each location along Aché each row represents Pittsburg the proportion of the Shona sample that made a ema) particular offer Au ai the right edge of the for example: the mode /the cai 1 _* lightly shaded horizontal most common offer is 0.15, tsimané grey bar gives the mean the secondary mode îs 0.25, Khazax offer for that group and the mean is 0.26 Torguud Mapuche Hadza ° ° e Machiguenga ° (} . . Quichua | 00 09 ° The graphs illustrates that: e nosociety conforms well to the predictions for Homo economicus; e there is much more variation in the offers made in the 15 small-scale societies than in the student populations of the advanced industrial countries: o the meanoffer varies from 26% (for the Machiguenga in Peru) to 58% (for the Lamelara in Indonesia); o the modal offer varies from 15% to 50%. e although we do not show any evidence here, it turns out that the rejection rates also vary quite considerably between the groups: o the rejection rate for offers below 20% / very low offers in industrial democracies varies from 40% to 60%, but rejection is quite rare in some groups. For example, the Machiguenga rejected only one offer even though 75% of the offers made were below 30%; o the rejection rate for very high offers are over 50% in some groups. For example, the Au and the Gnau in Papua New Guinea were equally likely to reject offers below or above 50%. The researchers found that individual level characteristics such asthe proposer's (or responder’s) sex, age, education, and wealth could not explain this large variation in behaviour between the different cultural groups. Instead, what mattered was how group-specific conditions such as social institutions or cultural fairness norms affected individual preferences or expectations. The researchers rank ordered the 15 societies along two dimensions: e the ‘payoffsto cooperation’ dimension, which refersto how important it isto cooperate with non-immediate kin in economic production: o at the low end of this dimension were the Machiguenga and Tsimané, whose members rarely engaged in cooperative production with individuals outside of the family; o atthe high endofthis dimension were the Lamelara, whose members hunted in 12 or more people atatime. 66 e the ‘market integration’ dimension, which refersto how much the groups relied on market exchange intheir everyday lives: o atthelowendofthis dimension were the Hadza, whose members rarely engaged in market activity; o atthehighendofthis dimension were the Orma, whose members often worked for wages and sold livestock. A statistical analysis revealed that higher values on both dimensions were associated with higher mean offers. One interpretation is that the individuals in these societies, rather than reason through the logic of the game, looked for similar situations in their everyday lives when they were faced withthe novel situations presented by the Ultimatum Game and the Dictator Game (“What familiar situation is this game like ?”), then, they acted in a way appropriate for this analogous situation: e consider the high offers and the rejections of both unfair and hyper-fair offers among the Au and Gnau in Papua New Guinea. This behaviour can be explained by the culture of gift giving. Providing expensive gifts is a signal of prestige and importance. At the same time, accepting large gifts commits one to reciprocate at some future time determined by the gift giver; e consider the low offers and the high rejections rates among the Hadza. This behaviour can be explained by the fact that Hadza hunters often try to avoid sharing their meat. One ethnographer calls this reluctance to share ‘tolerated theft’; e consider the Lamelara's tendency to divide the pie equally or to offer the respondent slightly more than a fair share. This can be explained by the fact that when a Lamelara whaling crew returns with a large catch, a designated person carefully divides the whale into predesignated parts allocated to the harpooner, crew members, and others participating inthe hunt, as well as the sailmaker, members of the hunters' corporate group, and other community members. What do these experiments have to do with culture and democracy? The results from these experiments suggest that culture might be considered a shared way of playing everyday games, that individual choices are shaped by the economic and social interactions of everyday life. If this istrue, then it seems reasonable to thinkthat some cultures might be less compatible with the game of democracy than it is in other societies. For example, the game of democracy often requires cooperation, competition, and compromise. Societies that already require this type of behaviour in their everyday games should find it easier to adopt and support democratic institutions. In contrast, societies in which individuals are engaged in games that do not encourage this type of behaviour will find it much harderto consolidate democracy. 8. DEMOCRATIC TRANSITIONS The number of independent countries in the world grew from 70 in 1946 to 192 in 2008. This increase was largely the result of the accelerated decolonization process forced upon European powers in the 1950s and 1960s and the breakup of the Soviet Union in the early 1990s. 67 e followinga big wave of strikes, in August 1988, the Polish government convened the Roundtable Talks with the main opposition group to help reach a compromise on how to deal with the growing economic and political problems. The result was the legalization of the independent trade union and nationwide elections in 1989, which produced the first non-Communist prime minister in Eastern Europe in 40 years; e three months after, the Triangular Table Talks in Hungary resulted in cautious moves toward easing censorship and legalizing an independent trade union. When the Soviet Union did not respond, further reforms were introduced: the Communist Party renamed itself the Socialist Party, the country's name was changed from the Hungarian People's Republic to the Republic of Hungary, and multiparty elections were planned for 1990. Although these changes in Eastern Europe were clearly significant, it should be noted that people at the time did not see them as signs of the imminent collapse of Communist control. The situation in East Germany began to change when Hungary decided to open its border with Austria in August 1989, thereby breaching the Iron Curtain for the first time. Thousands of East Germans fled to the West across Hungary s open border, leaving their possessions and relatives for freedom (although East Germans had always been relatively free to travel to other Communist countries in Eastern Europe, it had been all but impossible to get permission to travel to the West). An opposition group called ‘New Forum’ surfaced in East Germany demanding reform. Initially, the crowds chanted, “We want to leave!”; soon they began to chant, “We are staying here!”: protesters demanding reform and refusing to leave proved to be the real threat to the East German government. The early protests were small, but they soon began to grow as the East German government failed to successfully stopthe demonstrations. Gorbachev announced that the Soviet Union would not intervene in East Germany. Eventually, the East German government agreed on 9° November to remove all restrictions on travel to the West. The announcement of this decision on television led to tens of thousands of East Berliners rushing to the Berlin Wall. Despite brief attempts to create a nonsocialist East Germany, elections on 18°" March 1990 demonstrated that an overwhelming majority of East Germans wanted reunification with West Germany. Reunification finally took place on 3 October 1990, when the areas of the former German Democratic Republic were incorporated intothe Federal Republic of Germany. There are numerous other examples of bottom-up transition: the People Power Revolution that overthrew Marcos in the Philippines in 1986, the June Resistance which led to democratic elections in South Korea in 1987, the Velvet Revolution in Czechoslovakia in 1989, the overthrow of Ceaucescu in Romania in 1989. How can we explain these types of bottom-up transitions? Why did the collapse of communism in Eastern Europe occur in 1989 and not any earlier or any later? Why did Eastern Europe, which in retrospect seems to have been filled with extremely fragile Communist regimes, seem so stable before 1989? Why are revolutions so rare? Why are they so hard to predict? Collective action theory and tipping models will provide answers to some of these questions. Collective action theory Collective action theory focuses on forms of mass action / collective action, such as the protests in East Germany in 1989. Collective action refers to the pursuit of some objective by groups of individuals. Typically, the objective is some form of public good. A public good has two characteristics: e non-excludability + ifthe good is provided, everyone gets to enjoy it and nobody can be excluded from it; 70 e non-rivalry + if someone consumes the good, there is still just as much for everyone else to consume. We can think of democracy as a public good: it is non-excludable because anyone living in a democracy gets to enjoy living under democratic rule irrespective of whether they helped bring democracy about or whether they help to sustain it, and it is non-rivalrous because one person's enjoyment or consumption of democracy does not reduce the amount of democracy that others can consume. This might lead you to think that individuals who expect to benefit from a public good would be enthusiastic contributors to the provision of that good / that groups of individuals with common interests would act collectively to achieve those interests, at least as long as the benefit from the good outweighed the cost of providing it. The economist Mancur Olson has shown that there are reasons to doubt that this. The difficulty is the collective action problem / free-rider problem: individual members of a group often have little incentive to contribute to the provision of a public good that will benefit all members of the group. What possible difference could one person make to the success of a mass protest? Your individual decision to contribute to the public good or to participate in the pro-democracy protest is unlikely to be the decisive factor in determining whether the public good is provided or whether the protest is successful. Given that you are individually unlikely to influence the outcome of the protest, why pay the costs that come with participation? The costs of participation include time, possibly expense, and perhaps even loss of life. So the decision not to participate in a pro- democracy protest (or not to contribute to a public good more generally) is very appealing. If the pro-democracy rally fails, you will not have paid any costs of participation. If the pro-democracy rally succeeds, you can ‘free ride’ on the participation of others because everyone gets to benefit from the establishment of democracy irrespective of whether they participated or not. Imagine a group made up of N individuals (the entire East German population), of which K individuals (the number of pro-democracy protesters that are necessary to make the Communist government in East Germany back down and allow democracy to emerge) (where K < N) must contribute or participate for the public good to be provided (democracy). If democracy is achieved, then everyone receives a benefit B (where B > 0), irrespective of whetherthey participated or not. If you participate inthe pro-democracy rally, you must pay a cost C (where C > 0). If B>C (the benefit provided by the public good is higher than the individual cost of participating in the protest), everyone would have an incentive to contribute to the public good. If C > B (the individual cost of participatingin the protest is higher than the benefit provided by the public good), no one would ever have an incentive to contribute to the public good. Your decision of what to do will depend on your expectation about what other members of the group will do. Scenario 1 Scenario 2 Scenario 3 (Fewer than K-1 (Exactly K-1 (K or more participate) participate) participate) Participate -C B-C B-C Don't participate 0 o B Scenario 1: you expect that fewerthan K— 1 others will participate + it makes no sense for you to participate because your participation will not make the protest successful, and you will only end up payingthe cost of participation. You would be better off staying at home. The payoff of your best response in this scenario is 0. Scenario 2: you expect that exactly K — 1 participate + it makes sense for you to participate because your participation is decisive to turn an otherwise unsuccessful protest into a successful one. The payoff of your best response in this scenario is B—C. 71 Scenario 3: you expect that K or more others will participate + it makes no sense for you to participate because your participation is not necessary for a successful protest. You might as well stay at home and free ride on the successful participation of others without paying any costs. The payoff of your best response in this scenario is B. There are only two possible types of equilibria here: e equilibrium 1: no one participates + if no one participates, no one will want to individually deviate by participating because he or she will pay the cost of participating; e equilibrium 2: exactly K individuals do + if exactly K people participate, none of the participants will want to individually deviate by staying home because the rally will fail, and none of the non-participants will want to participate because the rally is already successful without them. Thus, for the collective action to succeed, exactly K individuals must believe that they, and only they, are likely to participate. This insight suggests that two factors in particular are crucial for determining the likely success of collective action: ® the difference between K and N: o ifK=N, then there is no incentive to free ride. In this situation, everyone knows that everyone must participate to obtain the public good; o if K<N, then there is an incentive to free ride. In this situation, everyone knows that they can obtain the public god without everyone's participation. However, if enough people free ride, then they will not obtain the public good. The larger is the difference between K and N, the greater is the incentive to free ride: - if Kisonlyslightlysmallerthan N, then most group members are still goingtothinkthat their participation will be crucial. As a result, the participation rates and, hence, the likelihood of success will remain relatively high; — if K is much smaller than N, then most group members are going to think that their participation will not be crucial. As a result, the participation rates and, hence, the likelihood of success will decline. * thesizeofN: o it influences the likelihood that you will think of yourself as criticalto the collective action. If hardly anyone else participates, thenthe pro-democracy protest is unlikelyto be successful. If lots of people participate, it is unlikely that your individual participation will make much of a difference. Given this, why participate? o it influences the ability of group members to monitor and punish free riders. The larger the group, the harder it isto monitor and punish those who do not participate inthe protest. The result is that larger groups tend to be characterized by higher levels of free riding and a lower likelihood that collective action will be successful. The relationship between group size and successful collective action suggests that small groups may be more effective than large groups because of the small group's increased ability to solve the free rider problem. Collective action theory provides a possible explanation for why public demonstrations of regime opposition are so rare in dictatorships more generally: the collective action problem may makes it difficult for the opposition to organize itself into a coherent force. call this a ‘revolutionary bandwagon': just as pro-democracy supporters falsify their preferences under dictatorship to avoid punishment, pro-dictatorship supporters will falsify their preferences under democracy for similar reasons. Supporters of the former Communist regime are likelyto lie about their true preferences and indicatethat they were long-standing opponents of the toppled government. In effect, the evidence from these people will make it seem as if the former dictatorship was much more unstable and unloved than it actually was. The Arab Spring Another revolutionary cascade occurred more recently in the Middle East and North Africa. On 17°" December 2010, a 26-year-old Tunisian man named Mohamed Bouazizi set himself on fire to protest against the local officials for confiscating the produce he was selling onthe streets and the electronic scales he used to weigh the fruit, and for harassing him. Bouazizi appealed to the local governor's office and, when ignored, stood in the traffic, and lit himself on fire. He died on 4° January 2011. Bouazizi's self-immolation was followed by mass protests which spread from his hometown to Tunis. On 15° January, after controlling the country for 23 years, President Ben Ali fled to Saudi Arabia, making him the first Arab leader in generations to leave office in response to public protests. Withthe diffusion of protests across national borders, across the MENA region, the Arab Spring was born: protesters in one country after another began to emulate the rebellious actions of protesters in other countries, and belief began to spread that the dictatorial regimes in the region could be toppled. e On 23" January, protests spread to Yemen where thousands took to the streets in support of Tawakul Karman, an activistwho was jailed after she called for an end to President Ali Abdullah Saleh's 32 -year control of the country. e On 25° January, organized protests, coordinated through social media, occurred across Egypt. President Mubarak resigned on 11° February and was tried and found guilty of corruption. e On 16°" February, protests erupted in Libya, a month after the leader Al-Qaddafi bemoaned the ouster of Ben Ali. Within a few weeks, mass protests had turned into a full-scale insurrection that was aided by NATO air strikes. Therefore, the removal of Al-Qaddafi was heavy reliant on foreign intervention. e By March 2011, significant protests had occurred in 17 countries throughout the MENA region. Information is central to the success of a revolutionary cascade. For example, protesters need to know about events in other countries. They also need to be able to pass on various pieces of information such as where and when to meet so that they can coordinate their protests successfully. Some commentators have referred to the Arab Spring as the Twitter Revolution’, suggesting that new developments in information technology, such as Twitter, Facebook, and YouTube, which enable information to be shared quickly with others, are empowering citizens, weakening authoritarian elites, and helping to make revolutionary cascades more likely. It is certainly true that protesters during the Arab Spring used these new technologies in various capacities. It is unclear, though, exactly how big a role these new technologies played inthe Arab Spring. Also, there is growing evidence that dictatorial regimes may well be able to use these new forms of social media to their own advantage: in 2009, China simply closed Twitter messaging sites and blocked access to websites that were reporting on the ethnic protests in Xinjiang province. Recognizing that people will often find alternative ways to reach the outside world if social network systems are simply shut down, some authoritarian regimes have adopted more ingenious strategies where they create state- controlled intranets and Internet providers that give people the illusion of access to the outside world: Iran has 75 established a ‘halal’ Internet that has been described as “a stilted alternative reality of government approved content on controlled national intranets”. In many ways, traditional off-line surveillance can be difficult and expensive for authoritarian regimes. The emergence of new technologies actually makes it easier and cheaper for them to monitor their citizens: e thegovernment in China often forces companies to log social network users’ Internet protocol (IP) addresses; e the Egyptian state security apparatus under Mubarak collected records of activists' text messages, e-mail exchanges, and Skype chats. 8.3 Top-down transition Some transitions result primarily from a policy of liberalization on the part of authoritarian elites themselves. This policy of liberalization is often designed to stabilize a dictatorship but sometimes inadvertently leads to democracy. Periods of liberalization have preceded numerous transitions to democracy throughout history: Brazil 1982 — 1985, Uruguay 1983 — 1984, Chile in the 1980s, Poland in 1989, East Germany in November 1989. “Myanmar's best-known political prisoner, the opposition leader, Aung San Suu Kyi, is detained at home in Yangon. Besides her, more than 2,100 other political prisoners are held, all in squalid and brutal conditions. Many are serving sentences of up to 65 years for peaceful political activities. Former detainees say that torture is routine, and that medical attention is often denied even when prisoners fall gravely ill. Under a “road-map for democracy”, Myanmar will this year vote in a “multiparty election”. Miss Suu Kyi's party, the National League for Democracy, is deciding whether to take part. It is a difficult choice. Joining in would add legitimacy to a process with a preordained outcome—army dominance. But no other sort of change is on offer. This week a court sentenced a Burmese-born American activist, Nyi Nyi Aung, to three years in prison for forging an identity card and violating immigration law. One League precondition to taking part inthe election isthe release of all political prisoners. The regime, however, seems intent only on adding to their number.” (The Economist, 11° February 2010) The transition game with complete information The prehistory of the top-down transition game / top-down transitions to democracy frequently result from a split between soft-liners and hard-liners in an authoritarian regime. This split is often caused by declining economic conditions or social unrest: e hard-linerstendto be satisfied with the political status quo; e soft-liners may preferto liberalize and broaden the social base of the dictatorship. A policy of liberalization entails a controlled opening of the political space and might include the formation of political parties, holding elections, writing a constitution, establishing a judiciary, opening a legislature, and so on. The goal is not to bring about democracy but to incorporate (or, at least, divide and control) various opposition groups into authoritarian institutions, to produce what we might call a ‘broadened dictatorship'. Many people believe that the liberalization process is a sign that these states are gradually moving toward democracy. It is this belief that often encourages some scholars to label these regimes as ‘mixed’, ‘hybrid’, ‘partial democracies', or ‘partly free’. However, empirical evidence shows that the liberalization process will not necessarily bring democracy, but it can significantly enhance the stability of dictatorial rule. Given the potential benefits of liberalization, why do authoritarian elites not always push for it? We will see that the soft-liners cannot guarantee that liberalization will successfully produce a broadened dictatorship. 76 Soft-liners Do nothing Open O1: Status quo Opposition Enter Organize 02: Broadened dictatorship Soft-liners Repress Democratize 03: Narrow dictatorship O5: Democratic transition or 04: Insurgency For some reason, the soft-liners have come to prominence and they have two options: do nothing > outcome 1: status quo; open up the regime / liberalize. Dr The democratic opposition has two options: enter + outcome 2: broadened dictatorship. The democratic opposition accepts the concessions offered by the authoritarian elites and enter the institutions of a broadened dictatorship. In this case, it essentially agrees to maintain the dictatorial rules of the game in return for entrance into the formal political sphere. The soft-liners would see this as a success; organize + the democratic opposition takes advantage of its new freedoms to further organize and mobilize against the dictatorship. The Polish Communist Party agreed to allow the formation ofthe independent trade union Solidarity in September 1980, and within just two weeks, Solidarity had 3 million members and was rapidly becoming a direct threat. The soft-liners would see this as a failure, and their position in the dictatorship is likely to be undermined. The soft-liners have two options: repress + outcome 3: narrow dictatorship, or outcome 4: insurgency. If the repression is successful, there will be a narrow dictatorship in which the soft-liners pay the consequence for having introduced a failed policy of liberalization and are replaced by hard-liners: the 1968 Prague Spring. If the repression is unsuccessful, there will be an insurgency; democratize + outcome 5: democratic transition. The soft-liners accept the demands and allow the emergence of truly democratic institutions: the Soviet Union under Gorbachev. Before we can determine what the players are likely to do inthe game, we need to know how the players value the possible outcomes. We provide a preference ordering for each player over the five possible outcomes: the soft-liners > broadened dictatorship > status quo > narrow dictatorship > democratic transition > insurgency; 77 We give from 5 to 1 fromthe most preferred to the least preferred payoff. These payoffs are called ‘cardinal payoffs' because they tell us how much more a player values one outcome compared with another. Outcome Description Soft-liners Opposition 01 Status quo 4 3 02 Broadened dictatorship 5 4 03 Narrow dictatorship 3 1 04 Insurgency 1 2 05 Democratic transition 2 s Now that we have specified the game, we can try to figure out what the players will do. We solve the game for the SPE using the method of backward induction: at the final two choice nodes, the soft-liners must choose whether to democratize or repress. As we saw earlier, the soft-liners will choose to repress when they believe the democratic opposition is weak, and they will choose to democratize when they believe the democratic opposition is strong; now the democratic opposition must choose whether to continue organizing or enter a broadened dictatorship. As we saw earlier, a weak opposition will enter and a strong opposition will organize; we must now determine whether the soft-liners will choose to do nothing or open up at their initial choice nodes. Unfortunately, the method of backward induction no longer works because the soft-liners do not know which side of the game tree they are on, so they do know for sure what their payoffs will be if they open up: o ifthe soft-liners open up and the opposition isweak, they can look down the game tree and see that the outcome will be a broadened dictatorship with a payoff of 5; o if the soft-liners open up and the opposition is strong, they can again look down the game tree and see that the outcome will be a democratic transition with a payoff of 2; o if the soft-liners do nothing, they will get a payoff of 4, whether the democratic opposition is weak or strong. What should the soft-liners do? The way political scientists approach this problem is to ask whether the payoff the soft-liners expect to get if they open up is greater than the payoff they know they will get if they do nothing /whetherthe expected payoff of opening up is bigger or smaller than the known payoff of doing nothing. To answer this question, it is necessary to calculate the soft-liners’ expected payoff from opening up. In general, the expected payoff of some choice is calculated by multiplying the payoff associated with each possible outcome by the probability that that outcome occurs and then summing these values. It tells the players what they could expect their average payoff to be if they were able to make their choice over and over again. If there are two possible outcomes associated with some choice, as is the case here, then the expected payoff is calculated as follows: Expected payoff (Choice) = (Probability Outcome 1 occurs x Payoff from Outcome 1) + (Probability Outcome 2 occurs x Payoff from Outcome 2) What is the expected payoff for the soft-liners if they open up? Expected payoff (Open) = (px 5)+[(1-p)x 2]=5p+2-2p=3p+2 80 Now that we know this, willthe soft-liners choose to do nothing and receive a payoff of 4, orwill they choose to open up with an expected payoff of 3p + 2? It should be immediately obvious that the soft-liners’ choice will depend on the value of p, that is, the probability with which the soft-liners believe the democratic opposition is weak. The soft-liners will choose to open up and liberalize when the expected payoff from opening up is greater than the payoff from doing nothing. This occurs when: Expected payoff (Open) > Payoff (Do Nothing) 3p+2>4 3p>2 p>23 As you can see, the soft-liners will choose to open up and liberalize if and only if they believe that the probability that the opposition is weak is greater than 2/3. This is referred to as the ‘critical probability. If they believe that the opposition is weak with a probability less than this, then the soft-liners will choose to do nothing and the outcome will be the status quo. Our transition game has two implications: e thefirstimplication is that we are likely to see institutionalized dictatorships (those with legislatures, political parties, elections, and so on) onlywhenthe soft-liners believe that the opposition is sufficiently weak. If they think that the opposition is strong, they will choose not to liberalize because they will not be able to repress the opposition if it looks like they are about to lose control of the liberalization process. But the game does not provide any firm predictions as to whether liberalization and institutionalization actually help the stability and survival of dictatorships. The effectiveness of the liberalization process will depend heavily on whether the beliefs of the authoritarian elites about the strength or weakness of the democratic opposition are correct. If a dictatorship establishes institutions such as elections and legislatures under the mistaken belief that the democratic opposition is weak, then these institutions may, in fact, accelerate the dictatorship's collapse; e the second implication is that it is possible to have soft-liners who would like to liberalize but who choose to do nothing because they are not sure that the democratic opposition is sufficiently weak that they could control the liberalization process if they start it. This suggests that some people living in dictatorships are actually living under more repressive conditions than need be the case. If the democratic opposition could somehow commit to not taking advantage of the liberalization process, then soft-liners might be willing to introduce some liberalizing reforms. The difficulty facing a weak opposition is that they cannot simply tell the soft-liners they are weak, while a strong opposition can say that they are weak in an attempt to trickthe soft-liners into introducing a liberalization process. Applying the transition game to Poland in 1989 In the late 1970s, the Polish economy was in a state of crisis, so the Communist government had borrowed money from the West to subsidize prices on things like food to keep its citizens satisfied. But from 1975 to 1981, Poland's foreign debt increased from $700 million to $23 billion. When the repayment of this huge debt began to come due, the leader of Poland's Communist Party, Edward Gierek, was forced to raise prices. This led to strikes that started in the Gdansk shipyards. The strikers presented 21 demands to the government, which included things like freedom of expression, a right to strike, access to the media, and better housing, but the most important one was for a trade union independent of Communist Party control. Unsure of whether the Polish army would actually fire on its own people, Gierek backed down, and the independent trade union, Solidarity, was formed in September 1980. Within two weeks of its formation, 3 million Poles had joined Solidarity under the leadership of an electrician by the name of Lech Walesa. By 1981, Solidarity had 10 million members and had come out in direct opposition to the Communist regime. 81 The Soviets were becoming increasingly worried about the situation in Poland, and there were rumours that they were drawing up invasion plans. By now, Gierek had been replaced by the hard-liner General Wojciech Jaruzelski,, who in December 1981 declared martial law. Overnight, thousands were arrested, the army occupied factories and smashed strikes, Solidarity was banned, and a military dictatorship was established. Martial law remained in place until 1983. This particular episode in Polish history illustrates just how a period of liberalization could get out of control and eventually produce a repressive backlash resulting in the coming to power of hard-liners like Jaruzelski. Despite attempts to solve the economic problems that plagued Poland, the economy did not significantly improve over the next few years. As a new wave of strikes threatened to get out of control and turn violent in mid-1988, in December 1988, the government agreed to convene a conference with the banned trade union Solidarity to see if they could hammer out a compromise. The Roundtable Talks took place between February and April 1989: it was agreed, among other things, that Solidarity would be legalized and that nationwide legislative elections would occur in June. The goal of this liberalization process was to defuse social unrest, co-opt the democratic opposition, and get Solidarity to lend its moral authority to the electoral process. As Jaruzelski put it, “the game is about absorbing the opposition into our system”. Like most elections in dictatorships, the proposed legislative elections in Poland were not entirely open. All 100 seats inthe Polish Senate were to be freely contested, but 65% of the seats inthe lower house, or Sejm, were reserved for the Communists and their allies. With 65% of the seats in the Sejm, the Communists could expect to appoint the prime minister; they were already guaranteed the presidency because there were to be no presidential elections. The Polish Communist Party expected to do quite well in the legislative elections and did not believe that Solidarity was strong enough to realistically challenge its hold on power. Just after introducing martial law in 1981, General Jaruzelski had established a special Center for Public Opinion Research (CBOS), designed to provide information about the level of support enjoyed by the regime among the public. Opinion polls since the mid-1980s had consistently shown that the Communist Party enjoyed more confidence than Solidarity. Given the beliefs that the Communist Party had about the strength of Solidarity, the election results of 4°" June 1989 came as a complete surprise. Solidarity won all 35% of the Sejm seats that it was allowed to contest and 99 of the 100 seats inthe Senate. The 65% of the guaranteed seats had ‘symbolically been divided between the Communist Party (37.6%) and the so-called ‘deaf and speechless’' puppet partiesthat were allied to it (26.9%). Afterthe election, two of these previously loyal parties joined forces with Solidarity to give it more than 50% of the Sejm seats. With this legislative majority, Solidarity was able to appoint the first non-Communist prime minister in Eastern Europe in forty years. In contrast to 1981, the Soviets now made it abundantly clear that they would not intervene in Polish affairs, which meant that the Communists could not feasibly break the agreement they had reached with Solidarity and reverse the liberalization process. The events in Poland had the effect of encouraging the democratic opposition in other East European countries to challenge their own Communist rulers. Within months, Communist control in Eastern Europe had essentially come to an end. As our Transition Game predicted, one of the key elements in Poland's democratic transition was the Communist Party's incorrect beliefs about the strength of Solidarity. The opinion surveys conducted by CBOS turned out to be poor indicators of the political support enjoyed by Solidarity. Although they were carefully conducted, they were deeply flawed in many ways. For example, roughly 30% of respondents refused to complete the surveys. To a large extent, this high refusal rate was the result of Solidarity supporters not wanting to interact with Communist institutions. Many of those who opposed Communist rule but who actually agreed to complete the surveys would have falsified their preferences out of fear anyway. How might this affect economic growth? Consider that potential investors (the rich) are always deciding whether to consume / spend their after-tax income or invest it. If they thinkthat the tax rate is too high, then they will prefer to consume rather than invest. In the absence of investment, economic growth can be expected to be slow. Although democracy might be expected to offer property owners protection against seizures by the state, the Meltzer-Richard model illustrates that democracy introduces the possibility that the poor will seize the property of the rich through a redistributive tax scheme. Why democracy might protect property rights There are, however, two criticisms of the Meltzer-Richard model. Both of these criticisms suggest that the model may overstate the extent to which democratic politics will lead to large-scale redistribution and, therefore, to growth- reducing disinvestment. 1 N Political participation in democracies is typically inversely related to income: poor people are less likely to vote than rich people. As a result, the tax rate in a democracy may not be that much higher than it would be in a dictatorship. The structural dependence of the state on capital suggests that capitalists have a veto over state policies, so their failure to invest at adequate levels can create major problems for the state managers. Why? A model by Przeworski starts with an economy comprising two groups: e groupP derives its income from profits + profit takers / the capitalists; e groupW derives it from wages > wage earners / the workers. P _ Technological possibility frontier Share of societal output going to profit takers ° Share of societal output going to wage earners For any given level of technological development in society, there is a maximum level of output that the economy is capable of producing. The maximum level of output can be achieved only if profit takers and wage earners efficiently invest their resources (their capital and labour) in the economy. They will efficiently invest only if they receive as much as the investments they make / are not heavily taxed. The technological possibility frontier (the downward-sloping line): e at P* the profittakers receive all of the output and the wage earners receive nothing; e at\W* the wage earners receive all ofthe output and the profit takers receive nothing; e inbetweenthese two points, there is a continuum of distributional possibilities. What the capitalist possibility frontier represents + individuals in a capitalist economy always have a choice about how to allocate their resources. For example, profit takers are always comparing what they could get 3 from investing their capital against the enjoyment they could get from simply consuming it. Similarly, wage earners are always comparing what they could get from an additional hour of work against the value they place on an additional hour of leisure. If profit takers or wage earners ever receive less than their entire return on their capital or labour investments / are heavily taxed then profit takers will start consuming more of their capital ratherthan investing it and wage earnerswill start taking more leisure time ratherthan working (because everything they earn goes to the state). As a result, society s resources will be underutilized, and there will be a decline in societal output away from what is technologically possible. Imagine a society in which the tax and transfer system has historically been controlled by an autocratic government representing the interests of profit takers where societal output is maximized by returning a larger share of society s product to the profittakers than to the wage earners (point M). Imagine that the autocratic government is replaced by a democratic one in which there are at least as many wage earners as profit takers. This change in regime would, presumably, put pressure on the government to use itstax and transfer system to redistribute wealth in a more egalitarian fashion (point E). So far, the predicted effect of a shift to democracy is essentiallythe same as that generated by the Meltzer- Richard model that we examined earlier: there will be pressure for redistribution. Because profit takers are no longer receiving the full value of their investments, they will begin to remove their capital assets from the economy, either by consuming them or by shifting them to a different economy (a process known as ‘capital flight’). As a result, societal output will decline from point E on the technological possibility frontier to point C on the capitalist possibility frontier. The distance between points E and C constitutes the cost in lost societal output that results from the redistributive policy. The Przeworski's model illustrates that there is a potential trade-off between equality and growth: increased equality can be obtained only by accepting reduced economic growth. In conclusion, structural Marxists argue that governments will recognize this trade-off and choose to promote growth over equality. Citizens’ incentives to consume rather than invest A second argument linking regime type and economic growth is based on the claim that democracies are expected to produce poor economic growth: it encourages workers to consume their assets immediately rather than invest them (we are now talking about investment by workers rather than investment by economic elites) because the poor cannot afford to direct their assets away from immediate consumption (they need to eat and pay the rent today) toward investment for the future; when workers can organize by forming political parties or trade unions, as they can in democracies, they have the ability to drive wages up, thereby reducing profits for business owners and, hence, overall investment; because workers can vote in democracies, the government also has incentives to direct money toward them and away from investment in the economy, thereby slowing economic growth even further. If democratic governments promote saving and investment too heavily, they will likely be voted out of office. In contrast, if dictators are future oriented, then they can use their power to force people to save, thereby launching economic growth. For this story we assume that: the poor have a higher propensity to consume than the rich + but do the poor really have a higher propensity to consume than the rich? e economicgrowth is primarily driven by capital investment + but is economic growth really primarily driven by capital investment? e dictatorscare more about the future than democratic leaders + but why would dictators care about the future more than democratic leaders? This second argument also fails to provide clear predictions about whether democracies or dictatorships will produce higher economic growth. Economic voting and political business cycles A related story about the effects of regimetype on economic outcomes argues that voterstendto reward officeholders when the economy is performing well. Consequently, incumbent politicians have an incentive to try to stimulate economic activity as elections approach. The result is a series of pre-election economic ‘booms’ followed by post- election ‘busts’. In this way, elections can increase volatility in macroeconomic indicators, such as gross domestic product, unemployment, and inflation. Evidence that political business cycles occur is mixed, but if they do occur, the increased volatility they cause is likely to be bad for the economy since risk-averse investors can be expected to respond to this volatility by consuming rather than investing. In this way, the classic political business cycle argument can be thought of as another example of how democracy might encourage current consumption overthe future benefits of investment /is expectedto produce poor economic growth. Dictatorial autonomy Athird argument linking regime type and economic growth is based on the claim that democratic leaders are subjected to many pressures from special interests because they can easily be voted out of office ifthey fail to retain the political and financial support of powerful interest groups. Some political scientists use the claim that dictators enjoy more autonomy from special interests than democratic leaders to argue that: e dictatorships are better for economic growth + they do not need to spend money in an inefficient mannerto keep certain groups happy. The assumption isthat the dictator chooses to promote economic growth, but why would the dictator necessarily choose to do this? e dictatorships are worse for growth + they will be able to behave in a predatory way because there is no one to constrain their behaviour. Economic elites will be less likely to invest because they will be worried that a predatory dictatorwill confiscate all of their wealth and profits. Given the lack of investment, economic growth is expected to be low. Political scientists who make this type of argument claim that economic growth will be lower in dictatorships than in democracies. The assumption is that democracies protect property rights (limit state predation) better than dictatorships, but we have already seen that this assumption is extremely hard to defend from both a theoretical and an empirical point of view. Empirical evidence What does the empirical evidence say? Do democracies or dictatorships produce higher economic growth in the real world? In an overview of 18 different studies examining the effect of regime type on economic growth, Przeworski and Limongi report that 8 find that dictatorships grow faster, 8 find that democracies grow faster, and 5 find that regime type has no effect on growth. They conclude that “we do not know whether democracy fosters or hinders economic growth... Our own hunch isthat politics does matter, but ‘regimes’ do not capture the relevant differences.” PART III. Varieties of democracy and dictatorship 10. VARIETIES OF DICTATORSHIP 10.1 A three-way classification of authoritarian regimes Cheibub, Gandhi, and Vreeland provide one common typology of authoritarian regimes. They suggest that a useful way to distinguish between dictatorships isin terms of how authoritarian rulers are removed from office, based on the idea that we can distinguish them based on the identity of their support coalitions, as all dictators, like their democratic counterparts, rely on the support of a coalition to stay in power. An implication of this is that when we see a dictator removed from power, we are likely to see him replaced by a defecting member of his own support coalition. As a result, we should frequently see dictators replaced by dictators of a similar type. Three things can happen when a dictator leaves office: 1. democratization may occur with the result that the authoritarian regime is replaced by a democratic regime; 2. the same authoritarian regime may survive but under new leadership; 3. the incumbent authoritarian regime may be replaced by a different type of authoritarian regime. What happened when 388 authoritarian leaders left office for reasons otherthan natural death between 1945 and 1996: Type of successor Type of current ZZZ??? dictator Monarchy Military Civilian Democrat Total Monarchy 11 6 4 1 22 Military 0 89 38 52 179 Civilian 2 53 103 29 187 Total 13 148 145 82 388 Dictatorial leaders are replaced by individuals from the same authoritarian regime about 50% of the time. This is the reason why we often speak not just of individual dictatorial leaders but also of dictatorial regimes: the survival of a dictatorial leader and the survival of a dictatorial regime are not the same thing. Monarchic, military, and civilian dictatorships from 1946 to 2008: a. Number of Dictatorships by Dictatorial Type ». Percentage of Dictatorships by Dictatorial Type fsi Percentage of Dictatorships Number of Dictatorships 2 N W a us a o 3 S 8 8 Ss 8 3 iu z ___+_ ‘ii SILELELELILESLIZAIAGLO do sITRIELELILERZIZIGEEAO 82 Ss a a a è & a $ È Ss 5 85 soesreairaaLd SR RR SR Year — Civilian Military -— Monarchy — Civilian Military -—-. Monarchy In 2008, there were 74 dictatorships around the world: of these, 38 (51.4%) were civilian, 24 (32.4%) were military, and 12 (16.2%) were monarchies. e The civilian form of dictatorship has always been the most common. e The heyday for military dictatorships was in the late 1970s when almost 40% of dictatorships were run by the military. There has been a significant decline in the number of military dictatorships since the end of the Cold War. e The number of monarchies has not changed a lot over time; this suggests that monarchies have been a particularly stable form of dictatorial regime. 1. Who is the effective head of government? 2. Does the effective head of government bear the title of “king” and have a hereditary successor or predecessor? __ Yes No MONARCHY 3. Isthe effective head of government a current or past member of the armed forces? Yes No MILITARY CIVILIAN Monarchic dictatorships A dictatorial monarchy is an autocracy in which the executive relies on family and kin networks to come and stay in power. Empirically, dictatorial monarchies are generally a stable form of authoritarian regime: less violence, monarchic leaders survive in office longer, more stable property rights, faster economic growth. But when a dictatorial monarchy does collapse, it isoften followed by periods of violence and the installation of an even more repressive authoritarian regime. Why are monarchic dictatorships so stable? As similar with other dictatorships, monarchs often seek to maintain the loyalty of their support coalition by allowing members of the royal family to colonize government posts that they can then use for their own material benefit. What is different in monarchies is that they have generated a political culture where a leader's promise to distribute rents to his support coalition is more credible than in other types of dictatorship. This monarchic culture rests on three things: e thereareclearrules asto who the insiders and outsiders are. In general, monarchies tend to depend on tightly knit family structures that are reinforced through intermarriage. These rules allow insiders to know that their privileged position inthe regime is relatively secure; e monarchies tend to have rules or norms that indicate exactly how regime rents (government posts) are to be shared among the various members of the royal family. Regime collapse threatens access to the political and economic rents they have been promised; e monarchies tend to have institutions that allow members of the royal family to monitor the actions of the monarch and enforce the norms regarding the distribution of regime rents. Military dictatorships A military dictatorship is an autocracy in which the executive relies on the armed forces to come to and stay in power. 9 Military rulers often portray “themselves as guardians of the national interest” (Cheibub, Gandhi, and Vreeland). Of course, it is not clear that military rulers actually have such altruistic motivations. Some scholars have argued that military coups are often more motivated by class conflict or corporate interests. The military tends to value discipline and cohesiveness, autonomy from civilian intervention, and military budgets large enough to attract recruits and buy weapons. Officers tend to participate in coups only when a government threatens the interests, or the very existence, of the military. Disagreements over, say, economic policy or the distribution of office benefits among senior officers can lead to factionalization. In these circumstances, many officers prefer to return to the barracks and allow elections rather than riskthe unity of the military by trying to clingto power. The value of the exit option (giving up power) is considerably higher for military dictatorships than for other forms of dictatorship. The fact thatthe military has allthe ‘guns’ means that it retains a crediblethreat to re-intervene in politics / it can step down from power with a greater sense of assurance that whoever wins the elections will still have to take account of the military's preferences due to the possibility of future coups. In many cases, the military will actually negotiate the handover of power to make sure that its interests are indeed protected. Empirically, military dictatorshipstend to have short durations and are more likely to end with negotiations as opposed to violence than other types of authoritarian regime. There is also some evidence that military dictatorships are more likely to leave behind competitive and democratic forms of government than other types of dictatorship. 1960-1990 1991-2004 Only 25% of the coups of the Cold War period were followed by competitive elections within 5 years. In contrast, 74%ofthe coups in the post-Cold War period were followed by competitive elections within 5 years. BM Election held less than 5 years after a military coup EI No elections within 5 years of a military coup What explainsthis dramatic difference between the Cold War and post-Cold War periods? Marinov and Goemans argue that foreign countries, particularly Western ones, exerted much less pressure on military dictatorships to hold elections during the Cold War period: they often preferred to support anti-Communist military juntas rather than encourage competitive elections that might produce left-leaning governments sympathetic to the Soviet Union, and in some cases, Western countries even helped the military with its coup (it is known that the Nixon administration and the CIA helped in the military overthrow of the left-leaning President Salvador Allende in Chile in 1973). According to them, the demise of the Soviet Union has enabled Western countries to push a more pro-democracy agenda. Since 1997, US presidents have been bound by the 1997 Congressional Act to suspend foreign aid to any recipient country that experiences a coup d'état. In line with their argument, Marinov and Goemans find that a country's dependence on Western aid increases the likelihood that a postcoup election will take place, but only in the post-Cold War period. This line of reasoning has led some to wonder whether coups might even be ‘good’ for democracy: e Thyne and Powell suggest that coups, by providing a ‘shock to the system, can create opportunities for liberalization that would not otherwise exist; 10 e theconcentration of office benefits inthe leader's faction meansthat a personalist dictator can more easily ride out periods of poor economic performance. Although ordinary citizens may well suffer in an economic downtumn, it is often the case that the dictator will retain sufficient resources to keep his support coalition satisfied; e the highly repressive nature of the security apparatus in a personalist dictatorship means that the probability of successfully overthrowing the regime is quite low; e membersofthe leader's faction ina personalist dictatorship have much less valuable exit optionsthan members of the regime party in a dominant-party dictatorship due to the fact that personalist dictators retain personal control of policy decisions and the selection of personnel. A consequence of this is that elites in personalist dictatorships often fight to the very end when their access to power is threatened. This helps to explain why personalist dictatorships are more likely to end in violence than other types of dictatorship. On the whole, then, personalist dictatorships tend to become unstable only: e whenthere is an economic catastrophe as opposed to a mild downturn; e whenthe security apparatus and military defect; e whenthe leader dies and the system of patronage based around him collapses. Dominant-party and personalist dictatorships are closed authoritarian regimes in which no opposition party is granted a legal space in the political arena. They can be contrasted with electoral authoritarian regimes. Electoral authoritarian regimes The increasing frequency with which elections are taking place in dictatorships has led some scholars to suggest that we are observing the emergence of a new type of dictatorship: the electoral authoritarianism. In an electoral authoritarian regime, leaders “hold elections and tolerate some pluralism and interparty competition, but at the same time violate minimal democratic norms so severely and systematically that it makes no sense to classify them as democracies” (Schedler). The extent to which competition is allowed in electoral authoritarian regimes varies: e hegemonicelectoral regimes > the leader's party routinely wins with overwhelming majorities, and there isno meaningful contestation. Elections are seen as forms of institutional window dressing with few political consequences (wrong view); e competitive authoritarian regimes + competition is more real and opposition parties are ableto win substantial minorities at election time. Elections are seen as a prelude to further shifts toward democratization (wrong view). Actually, the overall consensus is that authoritarian elections have very significant political consequences, and that dictators use them to help stabilize their rule. Elections can help dictators in at least three ways: e theycan help them co-opt elites, party members, or larger societal groups + elections can be used as an arena for patronage distribution and as a means of recruiting and rewarding local political elites; e theycan help dictators co-opt opposition groups, as well as divide and control them: o byallowingallopposition groupsto compete in elections, dictators provide accessto political office and some decision-making authority; o by allowing only some, and not all, opposition groups to compete in elections, dictators divides the opposition, thereby making it harder for opposition groups to overthrow them. 13 e theycan provide important information to the dictator + for example: o to identify their bases of support and opposition strongholds. In this way, dictators can use election results to identify which regions they should reward and which regions they should punish; o toevaluate the performance of their local officials. Dictatorial elites in China sometimes use low support at the polls in local elections to identify incompetent and poorly performing local officials. DICTATORSHIPS MONARCHY CIVILIAN MILITARY POLITICALLY ELECTORAL —» However, we have doubts about whether SLOSED AUTHORITARIAN these regimes truly represent a new type of | dictatorship. PERSONALISTIC. DOMINANT PARTY HEGEMONIC COMPETITIVE ELECTORAL AUTHORITARIAN 10.2 The two fundamental problems of authoritarian rule Political scientists have identified two fundamental problems of authoritarian rule that exist in all dictatorships: a) the problem of authoritarian power-sharing and b) the problem of authoritarian control. The problem of authoritarian power-sharing The problem of authoritarian power-sharing focuses on threats to dictatorial rule that come from within the authoritarian elite. Dictators never come to power on their own, and they rarely control enough resources to govern alone. Instead, dictators rely on a support coalition. When the dictator first comesto power, there is an implicit, and possibly explicit, agreement on how to share economic and political rents among the members of the support coalition. The problem is that in a dictatorship there is no independent third-party actor to enforce this power-sharing agreement. The only thing stopping the dictator from grabbing more power is the ability of the support coalition to replace him. When the threat to remove the dictator is credible, we have a contested dictatorship where power is shared between the dictator and his allies. Removing a dictator, perhaps via a coup, can be costly, though: ® ifthe coupfails, the coup-plotters are likelyto be imprisoned or killed; ® ifthe coup succeeds, it can leave lingering divisionsthat destabilize the authoritarian regime. However, the dictator is unlikely to publicly announce his intention to usurp power. A consequence of this isthat it can be difficult for the support coalition to distinguish between a situation in which the dictator is making a power grab and one in which he is allocating rents inthe pre-agreed manner. This can result in: e unnecessary coups where the support coalition attempts to remove a dictator who is following the original power-sharing agreement; ® missed opportunities where the support coalition fails to act against a dictator who is concentrating power in his own hands. This uncertainty creates incentives for the dictator “to try his luck and attempt to acquire power at their expense” (Svolik). If the dictator is able to make successive power grabs without being stopped, then it is possible for him to accumulate sufficient power that the support coalition no longer hasthe abilityto crediblythreaten to remove him. At 14 this point, the authoritarian regime has shifted from a contested dictatorship in which the dictator is constrained by his allies to a personalist dictatorship in which the dictator has effectively monopolized power. Dictators and their support coalitions clearly have an incentive to create a power-sharing agreement that allows the dictatorto stay in power and the support coalition to benefit from the dictator being in power. But the dictator cannot simply promise to abide by the power-sharing agreement because such a promise is not credible. The support coalition and the dictator must find a solution such that the support coalition receives credible information about the dictator's actions /tothe monitoring problem. A solution could be to create appropriate political institutions. In particular, decision-making bodies within legislatures or parties can provide a forum for exchanging information and deliberating about policy. These decision-making bodies provide the members of the support coalition with information about the actions of the dictator, making it less likely that the dictator can consolidate power without being called to account for his actions. Information on its own, though, is not sufficient to create a stable power-sharing arrangement. In addition, support coalitions must also have the ability to credibly punish the dictator if he reneges on their agreement. This raises the issue of whether support coalitions can overcome the collective action problems that arise when attempting to remove a dictator: e whenthedistribution of power between the dictator andthe support coalition is fairly even, support coalitions will find it easier to overcome collective action problems and punish rule-breaking dictators; e whenthe dictator is particularly powerful, there will be disagreement amongthe various factions inthe support coalition as to whether they should, or are even able to, move against the dictator. The problem of authoritarian control The problem of authoritarian control focuses on threats to dictatorial rule that come from the masses. What is to stop the masses from rising up and overthrowing the dictator? To a large extent, dictators have two distinct strategies for solving the problem of authoritarian control: e repressthe masses: o ontheone hand, repression can keep the masses under control; o onthe other hand, the dictator must rely on other actors to keep the masses under control. , T ypically the military does the actual repressing, and they may or may not share the same preferences as the dictator. By providing the military with the resources necessary to successfully repress the population, the dictator effectively empowers the military to act, if it wishes, against the dictator. Dictators face a trade-off: they can keep the military weak but run the riskthat they will be overthrown in a revolution if the masses rise up, orthey can maintain a strong military and expose themselves to threats from the military. How this trade- off is ultimately resolved is likelyto depend on the nature of societal opposition: o with smalland irregular protests, they can rely on the internal security forces and the secret police and the military can be kept weak, hence. The civilian government has control overthe military inthis type of situation; © with ongoing large-scale, organized, and armed opposition, they will have little choice but to rely on the military to stay in power. Both the dictator and the military recognize the crucial role that the military play in sustaining the authoritarian regime in this type of situation. Given this, the military has to be kept strong: in return for services rendered to the dictator, the military will demand policy concessions, large budgets to buy weapons and attract recruits, and autonomy from authoritarian control. 15