Are Parliamentary Systems Better

Are Parliamentary Systems Better? ABSTRACT The institutional differences between presidential and parliamentary rule are well known, yet the practical effects of these divergent constitutional arrangements within democratic polities have received scant attention. This paper employs a global dataset to test the relationship between a historical measure of parliamentary rule and fourteen indicators ranging across three policy areas: political development, economic development, and human development. We find a strong relationship between parliamentarism and good governance, particularly in the latter two policy areas. To the extent that these institutions influence the quality of governance, parliamentary systems may offer advantages over presidential systems of democratic rule. The constitutional structure of the executive has been a major research question in political science since the birth of the discipline. Beginning with Walter Bagehot and Woodrow Wilson, scholars of American and European politics have pondered the effect of unified or separated powers on campaigns and elections, on voter behavior, on the behavior of party elites, on the workings of the bureaucracy and the courts, and more generally, on the policymaking process. More recently, the structure of the executive has emerged as an important question within the field of comparative politics, a consequence of the rise of new democratic polities around the world. Thus, many of the same questions that arose initially in comparisons between the United States and the United Kingdom are now being raised in a global context. Underlying much of this work within American and comparative politics is the implicit normative question: are parliamentary or presidential systems superior? Which method of structuring the executive leads to better governance? It is difficult to read the literature on parliamentarism and presidentialism without catching a strong whiff of these underlying normative concerns (e.g., Haggard & McCubbins 2001; Lijphart 1992; Mainwaring & Shugart 1997; Skach 2005; Wilson & Schramm 1994). And it is right that they should be there, for the practical question facing constitution-makers around the world is one of institutional choice. Taking all possible causal effects into account, which system is preferable? We live in an age of constitutional reform, where core elements of the polity are subjected to continual criticism and not infrequent amendment (Reynolds 2002; Sartori 1994). If political science cannot offer guidance on this matter then the discipline is, in essence, passing the buck. Moreover, this fundamental normative question—if answerable—may lead to significant theoretical gains for the discipline. It is not simply a question of exigency. Of course, it may be that neither system is better, overall. Conceivably, the only 2 consequential difference between these two constitutional molds is that one system (presidentialism), by virtue of greater institutional fragmentation, offers greater resistance to change (Tsebelis 2000: 443). There may also be as many differences within as between each constitutional type. Indeed, presidentialism and parliamentarism are unified labels for variegated realities (Haggard & McCubbins 2001; Mainwaring 1993; Shugart & Carey 1992; Tsebelis 1995). Alternatively, parliamentarism and presidentialism each may demonstrate varying strengths and weaknesses along different policy dimensions (Weaver & Rockman 1993). Or it could be that the performance of the executive is contingent upon cultural, socioeconomic, and historical factors that vary from country to country and from period to period. If so, parliamentarism may be more appropriate in some contexts, and presidentialism in others. The modal position among political scientists appears to be that the issue is complicated, and no clear and consistent advantage can be found for either executive type. In any case, questions regarding the overall governance impact of these two fundamental constitutional forms seem well worth posing. Without taking a broad cut at these matters we cannot hope to answer the question of whether there might be distinct strengths and weaknesses along different policy dimensions or among different regions of the world or across different time-periods, or whether there is in fact some consistent result that holds across dimensions, measures, regions, and eras. Given its theoretical and normative importance, it is striking that the question of executive organization and its effect on the quality of governance has received so little systematic crossnational empirical analysis. There is, of course, plenty of speculation on these matters, a bit of formal modeling, and quite a number of case studies. In addition, many large-N crossnational studies have been conducted on what might be termed the “proximate” causal 3 effects of presidentialism and parliamentarism—e.g., on electoral behavior, legislative behavior, interbranch relationships, cabinet formation, the operation of the bureaucracy, the policymaking process, the organization of interests within civil society, and the size of the welfare state (e.g., Samuels 2007). Yet, none of these outcomes has unambiguous ramifications for the quality of governance. Indeed, the only outcome of obvious normative importance that has been subjected to extensive crossnational empirical analysis is the purported affinity between parliamentarism and democracy, a matter about which there seems to be very little scholarly agreement (contrast Cheibub 2007 and Stepan & Skach 1993). Our concern here is with issues of governance internal to democratic polities. That is, given a modicum of multi-party competition, what are the policy effects of different executive structures? This question has scarcely been broached.1 Granted, a number of writers have tested aggregated models of democratic governance that bring together a number of political institutions into a single variable, e.g., the Consensus model (Lijphart 1999), the Checks index (Beck, Demirguc-Kunt & Levine 2000), the Political Constraints index (Henisz 2000), and the Centripetalism index (Gerring, Thacker & Moreno 2005). Our interest here is more narrowly targeted on the construction of the executive and its possible role in achieving good governance. In attempting to judge this matter we observe three broad policy areas—political 1 Foweraker & Landman (2003), whose primary concern is with the quality of democracy in parliamentary and presidential systems, also touch on the provision of property rights across regime types. Although they find no statistically significant difference between them along this dimension of economic policy, the regression analysis relies on a very small sample (N=40) that is potentially biased and subject to a host of possible specification errors (due to the paucity of control variables included in the model and lack of robustness tests). 4 development, economic development, and human development—comprising a total of fourteen specific outcomes understood as indicators of good or bad governance across these various dimensions. The first section of the paper presents the theoretical debates surrounding the efficacy of presidential and parliamentary systems. These debates generate the competing hypotheses tested in the empirical section of the paper. The paper concludes with a discussion of the results and implications of the research. THEORETICAL DEBATES The debate over presidentialism and parliamentarism has a long history, and the terms of this debate have changed little over the past century. The debate persists despite the fact that the proximate effects of different executive structures are generally accepted. Ceteris paribus, it is understood that parliamentary systems foster stronger political parties, more centralized and party-aligned interest groups, a more centralized decision-making process, and more centralized and hierarchical administrative structures. These differences are aptly, if somewhat ambiguously, summarized by the concept of veto points (Tsebelis 1995). Separate powers introduce an additional veto point into the political process, fragmenting politics at all levels. The effect of this concentration or dispersion of power is to create a contrast between systems that are decisive and those that are resolute (Cox & McCubbins 2001: 27). Parliamentary systems are decisive, presidential systems resolute. In addition, parliamentarism fosters a style of politics and policymaking that is probably more institutionalized, centered as it is on political parties, while presidentialism fosters a more personalized and free-floating style of leadership centered on individual politicians and smaller, less established organizational entities. Yet, none of these proximate causal effects has clear-cut implications for the quality of governance in a polity. Partisans of separate powers prize the diffusion of power, while 5 parliamentarists prize its unification. Thus, most of the arguments that we review here may be understood as derivative of a single centralist/decentralist dimension. Consider, first, the question of party organization. There is general agreement that the existence of an independently elected executive with strong policymaking prerogatives weakens party cohesion (Shugart 1998). Members of the same party are less likely to vote together in the legislature and party allegiances are weaker in the electorate. But there is intense disagreement over whether strong (cohesive) political parties foster good governance. Supporters of the Responsible Party Government model view strong political parties as the linchpin of democratic accountability and effective governance (APSA 1950). Critics of this ideal point out that undue partisanship may blind voters and legislators to the public interest and may prevent them from reaching compromise. Good public policy, not the interests of a party, should guide behavior in the ballot box and on the floor of the assembly (Herring 1940). Consider, next, the question of interest organization. A separate powers system is generally thought to encourage the formation of a highly fragmented, non-party aligned (“independent”) interest group community. Separate powers is thus a key feature of interest group pluralism, while parliamentarism is a key feature of corporatism (Wilson 1992). Pluralism may have positive ramifications for the quality of governance insofar as a greater diversity of views will be expressed, each group will retain a high degree of independence from political parties and from the state, and important groups will exercise an informal veto power over questionable legislation (Truman 1951). It may have negative ramifications for governance to the extent that groups represent narrow (“special”) interests, are able to veto legislation that serves the general interest, and defect rather than cooperate with other political actors (McConnell 1953; Olson 1982). 6 Consider, third, the question of bureaucratic organization. Lines of bureaucratic accountability are more complicated in a separate powers system since bureaucrats have two masters: their nominal chief executive and the controllers of the budget (Huber & Shipan 2002; Moe & Caldwell 1994). However, the effects of this dispersion of authority are disputed. Defenders argue that bureaucratic accountability is assured through multiple principals, semiindependent agencies carrying strict and highly specific mandates, and overlapping jurisdictions, such that bureaucrats check and balance each other and offer comparative benchmarks that make the task of policy effectiveness easier to assess (Breton 1996). For example, the price of bribes may be driven down by the opportunity to forum-shop among various agencies (Rose-Ackerman 1978). Opponents of bureaucratic fragmentation point out that it hinders cooperation among agencies, is less efficient, leads to problems of micro-management, and, by blurring lines of responsibility, erodes principal-agent accountability (Moe & Caldwell 1994). Consider, fourth, the question of transparency, presumably a key element of democratic accountability. Advocates of separate powers underline the fact that a direct vote for the chief executive leads to a transparent relationship between elections and electoral outcomes. Advocates of parliamentarism point out that post-election coalition negotiations affect only those circumstances in which no single party is able to gain a majority of the legislature. They also point out that these coalition negotiations are usually fairly predictable (it is rare for ideologically distant parties to coalesce); often, they are explicitly advertised such that two coalitions-inwaiting face off in the general election. In either instance, they scarcely attenuate the level of accountability that exists between the voter and his or her chosen party. Consider, fifth, the question of information. Advocates of presidentialism rightly point out that the division of power between executive and legislature creates an information-rich 7 environment. Each branch has an incentive, and the requisite constitutional authority, to investigate the other branch; each also has an incentive to publicize (either by formal proclamation or informal leaks) information favorable to the achievement of their political power and policy preferences. Thus, wherever there is disagreement between the two branches, or some important actors within these rather fragmented bodies, we can expect that the public will learn a good deal about policies and processes of government (Persson, Roland & Tabellini 1997). Advocates of parliamentarism counter that more information does not always lead to better governance. There is such a thing as too much information, or worse, misleading information. If members of the two branches are constantly attacking each other, engaged in vituperative “smear” campaigns, then the resulting information will not serve as a useful check against bad policies. It will, instead, enhance citizen apathy and alienation (Cox and Kernell 1991). Parliamentarists point out that plenty of information is publicly available in the normal workings of a democratic polity with a free press. The more important issue is how that information is framed and disseminated. Consider, sixth, the vexed question of electoral accountability. Advocates of separate powers see greater accountability at the local level—between constituents and their individual representative, whose positions and legislative record are distinct from his party and from the major-party presidential candidates. Advocates of parliamentarism see greater accountability at the national level—between constituents and their political party. In the process of governing, separate powers advocates see accountability arising from the diffusion of authority among multiple independent bodies, who are thereby able to effectively monitor, and if necessary block, each others’ actions. For them, the process of accountability is continuous—exercised through multiple bodies, each “checking and balancing” the other 8 (Laffont & Meleu 2001; Persson, Roland & Tabellini 1997). Advocates of parliamentarism, on the other hand, see accountability arising from the concentration of authority in a single set of hands (the ruling party or coalition and its leadership). For them, accountability is a retrospective process of reward or punishment by which principals (voters) exact accountability from agents (elected officials). Whether the principal-agent relationship concerns voters and representatives or representatives and bureaucrats, the chief point at issue is the relative clarity of responsibility and the prospect of ex post facto reward and punishment. From this perspective, accountability is rendered more tenuous, and the tendency to play “blame games” more tempting, whenever responsibility for governance is diffused across multiple bodies (Fiorina 1980; Linz 1994). On the other hand, strict accountability between elected officials and voters may not lead to better governance outcomes. Insofar as voters’ demands are unrealistic, short-sighted, or simply wrong-headed, it may be useful for government officials to find ways to mask their responsibility for unpopular measures. Thus, measures that impose costs on powerful constituencies with less visible benefits to the general public (or to other constituencies)—e.g., the closure of unneeded military bases, the lifting of trade barriers on protected sectors, or budget-balancing initiatives (raising revenue or reducing expenditures)—may benefit from the plausible deniability of responsibility by government actors. Since authority is more divided in a separate powers regime, it may be easier to engineer this gerrymandering for the public good (Pierson & Weaver 1993). Consider, seventh, the question of political change. Advocates of separate powers often emphasize the virtues of political stability. The primary task of politics is to assure a stable political environment so that the behavior of political actors is predictable from one election to 9 the next. And only where institutions are constitutionally separate can the government attain credible commitment to a set of policies (Henisz 2000; Keefer & Stasavage 2003). By contrast, advocates of parliamentary rule emphasize the problem of the status quo. For them, the chief political problem is to adapt to changing demands and changing circumstances. Political and economic sclerosis is the problem; flexible government, capable of overcoming strong vested interests and exercising leadership, is the solution (APSA 1950; Olson 1982). The same contrasts also apply to the short term. Presidential systems are rigid, in the sense that, absent an impeachment, there is no way to remove a sitting president in between elections (usually held at fixed intervals) (Linz 1990; Shugart & Carey 1992: 28-30). This lends the separate powers system a high degree of short-term stability. Parliamentary systems are more flexible since the prime minister, and the ruling coalition at-large, may be removed at any time by parliamentary vote and elections may also be called at any time. Stability in the executive is a virtue insofar as it allows leaders to credibly commit to policies and to stay the course; it is a vice insofar as presidents may lose legitimacy or pursue wrongheaded policies (Linz 1990). The question of political change is differently framed if viewed from the lens of policy innovation, the proposal and consideration of new policy ideas. Arguably, a separate powers system, because it assures more points of access and greater independence on the part of politicians (each of whom seeks to carve out his or her own turf on the national stage), provides a breeding ground for new ideas. Every interest group and every candidate is a potential policy entrepreneur. On the other hand, there is no guarantee that any of these ideas is worthwhile or politically viable. Many new ideas are called forth, but few are chosen. Consider, eighth, the question of institutionalization. Parliamentarism fosters a highly 10 predictable, institutionalized form of politics and policymaking in which participants are part of the establishment. Presidentialism, by contrast, fosters a more personalized form of political behavior in which presidents, legislators, interest group leaders, and even bureaucrats all enjoy a degree of independence from the institutions of which they are nominally members (Linz 1990; Lowi 1985). Each may play the role of policy entrepreneur, as we have noted. Each may play an oppositional role. Under the influence of the media, this political structure may lead to greater political conflict than would be expected in a parliamentary system, where negotiated settlements (often behind closed doors) are the norm. Although the institutionalization of conflict is, in general, a good thing, some may question whether parliamentarism fosters too much of a good thing. Achieving good governance is often a matter of achieving a proper balance between various attributes. Consider, ninth, the question of contestation and consensus. In a separate powers system, conflict is endemic and continual. Each branch is assumed to represent a somewhat different constituency, or to represent the same constituency in different ways. Yet, because a higher threshold of consensus is necessary in order to agree on any policy measure, it might be said that consensus is mandated by a separate powers constitution. In a parliamentary system, by contrast, power is temporarily monopolized by a single party or coalition. Other groups may voice their opposition, but they have no formal mechanism by which they might affect policy outcomes. Consequential conflict is thus episodic, occurring during elections but not in between. Thus, it might be said that a separate powers system is one where all parties engage in policymaking at the same time (with greater or lesser influence on policy results), while a parliamentary system is one in which parties take turns (dependent upon electoral results and post-election coalition agreements). 11 In which sort of system is the public interest more likely to be served? Cox and McCubbins (2001: 28; see also Linz 1994) stipulate that “the greater the number of effective vetoes, the more private regarding will be the policies enacted.” This, they argue, is a consequence of the bargaining process among veto holders, “where each veto player will be able to demand, and receive, side payments in the form of narrowly targeted policies. Thus, when the effective number of vetoes is great, even broad public policy will be packaged as a set of individual projects, or it will be packaged with narrowly targeted programs.” However, another sort of outcome is also possible. If a single policy can be found that offers a Pareto improvement for all veto holders, then no bargaining is necessary (Buchanan & Tullock 1962; Persson, Roland & Tabellini 1997). Indeed, the multiplication of veto holders also grants vetoes to persons who may wield them against special interest legislation or unwise policies. Veto power may prevent a variety of pathologies, including the tyranny of the majority, populist-inspired legislation, and socially inefficient redistributive coalitions (Riker 1982). Consider, finally, the question of strength or energy (“decisiveness”) in the executive. To some, this is a sign of danger, of corruption: an invitation to the abuse of power. To others, however, energy in the executive is a sign of state capacity. From this perspective, government must be empowered to do the right thing; otherwise, it will do nothing at all, or worse, different things at the same time. Cox and McCubbins (2001: 29) point to the serious problem of indecisiveness encountered in many presidential regimes, particularly when operating under conditions of divided government. These problems include “‘institutional warfare’ of varying intensity, unilateralism (where the executive and the legislature attempt to circumvent each other in implementing policy), various forms of gridlock, greater fiscal pork and rents, and a tendency toward larger budget deficits.” Those concerned about problems of indecisiveness look more 12 favorably on parliament’s concentration of political power in the hands of a single party or coalition. In sum, academics and policymakers working in this area seem to agree on one thing, and one thing only: constitutional structures matter. They disagree on how, specifically, constitutional structures such as the separation of powers affect the quality of governance in a democracy. On this central question there is no consensus, even though the proximate effects of presidentialism and parliamentarism are well understood and generally agreed upon. Disagreements hinge upon basic assumptions about the state, the citizenry, the workings of the marketplace, and the nature of politics. Not surprisingly, these opposing theoretical suppositions are longstanding. Arguments by Montesquieu, Madison, and their contemporaries (Vile 1967/1998) are echoed by advocates of separate powers in the present era. Arguments by nineteenth century parliamentarists such as Walter Bagehot, Woodrow Wilson, and Henry Jones Ford are reiterated by advocates of Responsible Party Government in the twentieth century (APSA 1950). It is testament to the recalcitrance of a theoretical debate that it fails to progress. Each side remains entrenched; each argument retains plausibility. Indeed, there are strong theoretical grounds justifying both sides of the argument. Each virtue of presidentialism (parliamentarism) can also be seen as a vice, depending upon one’s theoretical lens. They are mirror images of each other. Alternatively, arguments for each side may cancel each other out—the null hypothesis. This is all that we can safely conclude on the basis of our brief review of the literature. Is it possible to move forward toward a coherent—i.e., unified—theory for why presidentialism or parliamentarism might lead to stronger governance? This is, of course, the 13 ultimate objective. However, the reader will observe that the set of possible and plausible causal mechanisms listed in the foregoing pages is very large, and equally diverse. A single conceptual schema may not be able to capture these realities. Even if it could, it would likely become so overburdened with caveats and conjectures that it would serve as little more than a laundry list of possibilities. Alternatively, it is always possible to construct a coherent theory by the simple expedient of reductionism, i.e., by ignoring issues that do not fit neatly into the specified theory. But this is hardly convincing, given the strength of these manifold claims in the present instance. It is also possible to create the appearance of coherence by maintaining a sufficiently high level of abstraction. Yet, such models tend to resist empirical testing; they may be true or they may be false, but we can never know for sure (Green and Shapiro 1994). In the absence of a genuinely coherent, operational, and plausible theory of how democratic institutions work, we propose here to reverse the conventional sequence of scientific inquiry. Rather than theorizing first, then testing, we propose to test first, in the hopes that fruitful theorizing will follow. It is a well accepted feature of scientific progress that new theories often arise only when faced with a set of empirical findings that are striking enough to warrant theoretical exploration (Lakatos 1978). Findings thus pave the way for theory. And it is a procedure followed in some recent work. For example, Persson & Tabellini’s (2003) wideranging study of the economic effects of constitutions offers only a thin patina of theory, followed by an extensive empirical tour. This influential book is aptly subtitled, “What do the Data Say?” We adopt a similar approach in the present context. Fortuitously, the question of executive organization and governance is somewhat more tractable empirically than it is theoretically. We can measure inputs and outputs in the governance equation, even if we cannot test specific causal mechanisms. We therefore view the 14 present study as a first step towards the formulation of a new theoretical framework. In the concluding section, we offer some thoughts on how the multiple threads of a potentially unifying theory might be knit together. CLASSIFYING EXECUTIVE TYPES With the spread of democracy around the world in the postwar era comparativists face daunting challenges of definition and classification. What is a “parliamentary” system, how does it differ from a “presidential” one, and how might we classify the various intermediate types? Although we must impose a single definition on this complicated matter, it is important not to exaggerate the differences among these various classifications, which usually affect only a handful of borderline cases (mostly falling in the “semi-presidential” zone). Thus, any results, while to some extent contingent upon matters of classification, are likely to be robust across alternative classifications. We understand parliamentarism as a system of government in which the executive (the prime minister and cabinet: collectively, ‘the government’) is chosen by, and responsible to, an elective body (the legislature), thus creating a single locus of sovereignty at the national level. Presidentialism, its contrary, is a system where policymaking power is divided between two separately elected bodies, the legislature and the president. The president’s selection is usually by direct popular election, though it may be filtered through an electoral college (as in the United States), and the rules pertaining to victory (i.e., by relative or absolute majority) vary from country to country. His (or her) tenure cannot be foreshortened by parliament except in cases of gross malfeasance. S/he is actively engaged in the making of public policy, and in this sense plays a political (i.e., partisan) role. Between these polar types we find various admixtures, known generically as semi- 15 presidential polities (Bahro, Bayerlein & Veser 1998; Elgie 1999; Roper 2002; Skach 2005). Thus, we conceptualize the parliamentary/presidential distinction as a continuum with two dimensions: a) the degree of separation (independence) between president and parliament (unity = parliamentary, separation = presidential), and, if there is any separation at all, b) the relative power of the two players (the more power the president possesses the more presidential is the resulting system). We capture this complex reality with a three-part coding scheme: 0 = presidential, 1 = semi-presidential, 2 = parliamentary. As a general name for this three-part coding scheme we employ the term parliamentarism. However, since our interest is in the causal effect of parliamentarism/presidentialism (hereafter, parliamentarism), we must concern ourselves not only with a country’s current status but also its institutional history. How many years has it retained a parliamentary or presidential system? We suppose that it takes time for institutions to exert an appreciable effect on governance outcomes. A country switching from a presidential to parliamentary system (or establishing a parliamentary system in a newly democratic or independent setting) should not expect to see immediate, dramatic changes in the quality of governance. Instead, these effects are likely to cumulate over time as the new institutional rules begin to affect actions and expectations (Gerring, Thacker & Moreno 2005). To represent this historical component empirically we create a new variable that draws upon the annual (“raw”) scores for parliamentarism described above. We assume that history matters, but that recent history matters more. Thus, we calculate a moving, weighted sum of the annual scores of each of our underlying explanatory factors, beginning in 1901 and ending in the observation year. Weights capture long-term historical patterns while giving greater weight to more recent years. A country’s score in 1996 is the weighted sum of its scores from 1901 to 16 1996. Its score in 1998 is the weighted sum of its scores from 1901 to 1998, and so on. Country-years figure in this weighted summation process so long as a country surpasses a minimum threshold of democracy during that year. (Recall that we are concerned with democratic governance; parliamentarism has quite a different meaning in an authoritarian setting, if it has meaning at all.) We employ a relatively low threshold of democracy because we wish to include as many plausible cases as possible in our analysis and because we believe that the distinct dynamics instilled by parliamentarism or presidentialism should be in evidence in any situation where there is a modicum of multi-party competition. Thus, we include a countryyear in the following analysis so long as it obtains a score greater than zero—on a scale ranging from -10 to 10—on the Polity2 measure of democracy (drawn from the Polity IV dataset [Marshall & Jaggers 2002]). DEPENDENT VARIABLES In order to test the possible causal effects of parliamentarism on the quality of governance, we must identify a set of outcome indicators that are measurable, valid crossnationally, and reasonably clear in their normative implications (good or bad relative to the public interest). Length limitations preclude an extensive discussion of our choice of outcome variables (see Gerring & Thacker 2008). Suffice to say, we include as wide an array of governance outcomes as possible: those that are survey-based (“soft”) and those that are directly measurable (“hard”), those that measure processes (e.g., the workings of the bureaucracy) as well as those that measure policy effort or policy outcomes. There is no single best way to measure the quality of governance; our recourse is to extend the empirical purview to a wide range of plausible measures and to discern lessons from the broader patterns observed. It is especially important to take a wide cut at this material for the simple reason that any 17 individual result may be spurious. Regardless of the care one exercises in the choice of control variables and tests for robustness, it is still possible that empirical correlations with a given outcome variable may be driven by unmeasured factors (a problem of simultaneity or countryspecific effects) or measurement error. Cross-country regression is a sensitive empirical tool, and few results are robust across all specifications. However, if one observes a fairly consistent pattern across a wide range of associated variables, greater confidence is warranted. Outcome measures number fourteen in total and fall loosely into three policy areas: political development, economic development, and human development. Indicators of political development include two measures of corruption control (one from Political Risk Services [Howell 1998] and another from the World Bank’s governance indicators [Kaufmann, Kraay & Mastruzzi 2003]). We also employ a measure of bureaucratic quality (from PRS). Finally, we include three additional World Bank indicators: government effectiveness, political stability, and rule of law (Kaufmann, Kraay & Mastruzzi 2003). In each case, higher scores suggest better governance in the area of political development. We include five indicators of economic development. The number of telephone mainlines per 1000 inhabitants captures an important dimension of economic infrastructure and communication (World Bank 2003). Import duties as a percentage of imports provide a measure of trade protection, a negative governance outcome (World Bank 2003). Trade openness, measured by the value of total trade divided by gross domestic product, is an outcome-based measure of a country’s trade policy (World Bank 2003). To measure the overall investment climate, we use Euromoney’s country risk ratings, which are based on polls of economists and political analysts and supplemented by quantitative data such as debt ratios and access to capital markets (Euromoney 2004). Finally, our broadest economic measure captures the level of 18 prosperity within a country, as measured by real per capita GDP (World Bank 2003). Except in the case of import duties, higher values indicate better economic governance. Three variables capture the level of human development in societies around the world. The infant mortality rate (per 1000 live births, age 0 to 1) is a critical health and quality-of-life indicator (data from World Bank 2003). Encapsulating the entire life cycle are estimates of life expectancy (World Bank 2003). Illiteracy measures the percentage of the population that is unable to read and write a sentence about daily life in their native language (World Bank 2003). Lower scores for infant mortality and illiteracy and higher scores for life expectancy indicate better governance performance. EMPIRICAL STRATEGY Our method of analysis is largely contingent upon the nature of the evidence at hand. Where data are not comparable over time, as with the World Bank governance indicators, we employ a between-effects estimator. This approach averages the results of a series of cross sections, with data drawn from four adjacent time periods (measured every two years from 1996 to 2002). Where comparable time-series data are available, we employ a time-series—cross-section (TSCS) format. Although this sort of analysis offers distinct advantages over simple crosssections, it is subject to simultaneous spatial and temporal difficulties. We cannot employ a unitbased fixed-effect research design to address spatial issues, such as unobserved heterogeneity, because our causal variable, Parliamentarism, does not vary sufficiently from year to year. Its movement through time is “sluggish” (Wilson & Butler 2003; see also Beck, Demirguc-Kunt & Levine 2000: 285; Beck & Katz 2001, 492-93). We do, however, employ a set of regional fixed effects and a geographically weighted version of the dependent variable (see below) to help mitigate sample heterogeneity. With respect to temporal issues, we employ a statistical 19 correction for first-order autocorrelation, and a time-trend variable to control for possibly spurious correlations between heavily trended dependent and independent variables. The literature on governance suggests a wide array of political, economic, geographic, demographic, and cultural factors as plausible control variables. While we cannot measure all of these influences easily in a global setting, we include all the standard controls plus some not-sostandard ones that, we surmise, may also play an important role. A country’s regime history is likely to influence the quality of governance (Gerring et al. 2005). Thus, we include a variable that captures the cumulative democratic history of each country over the course of the twentieth century. A country’s level of economic development is also likely to be strongly associated with good governance. We control for this in standard fashion, by including GDP per capita (in logarithmic form). We measure this once, in 1960, rather than on an annual basis, to alleviate concerns about endogeneity between GDP and our various outcomes (World Bank 2003). (Results obtained using contemporaneous annual scores for per capita GDP, rather than the base year, yield the same substantive conclusions as those presented here.) We include regional controls for Africa, Asia and Latin America/Caribbean to capture whatever cultural, demographic, geographic, historical, and socioeconomic factors might be common to those regions. Dummy variables for Africa and Latin America/Caribbean are expected to reflect lower levels of governance in those regions compared to others, while expectations for Asia are somewhat mixed but generally more favorable. We anticipate that a significant period of socialist rule (LaPorta et al. 1999) has negative effects on our political and economic development indicators (with the exception of telephone mainlines, with which we anticipate a positive relationship) and positive effects on human development. Having an English legal origin is often thought to promote good governance (LaPorta et al. 1999). Other 20 colonial systems are thought to have been uniformly bad for governance outcomes in the developing world and hence are encapsulated in this variable. To the extent that countries farther from the equator have better governance, latitude (absolute value, scaled to 0-1, logarithm, LaPorta et al. 1999) should correlate with better outcomes. Expectations for ethnic (and linguistic) fractionalization (Alesina et al. 2003) are more tentative, but heterogeneity is generally expected to hamper the quality of governance in a country. To the extent that having a large population (total population, logarithm, World Bank 2003) makes certain governmental tasks more difficult, population might be expected to diminish governance capacity (except in the case of investment rating, where expectations are ambiguous because population may serve as a proxy for market size). Distance (in thousands of km) from the nearest financial center (Tokyo, New York, or London) is intended to capture the negative impact of geographic distance from the ‘cores’ of the international economy. Oil (millions of barrels per day per capita) and diamond (rescaled to billions of metric carats per year per capita) production levels capture the ‘resource curse’ (Humphreys 2005). Yet, these resources also provide sources of revenue and wealth. As such, expectations are mixed. We also include a control variable that measures the average value of the dependent variable across all countries, weighted by the inverse of the geographic distance (in kilometers) of each country from the country in question. (In the case of GDP per capita, we weight the average value of the dependent variable by each country’s share of trade with the observed country, rather than by the inverse of the geographic distance between the countries.) Countries lying close to one another may display similar values for extraneous reasons (culture, geography, diffusion, and so forth). Thus, we anticipate a positive sign for this variable. The inclusion of this variable in all regressions should help minimize possible spatial autocorrelation in the 21 sample. We employ additional variables in selective regressions, as appropriate. Thus, a measure of Protestantism is included in the analysis of political development and Islam in the analysis of human development outcomes (both are measured as a percent of the total population). Prior research suggests that a Protestant heritage may improve state capacity (Gerring & Thacker 2004), while having a large Muslim population may impede human development (Moon 1991). Linguistic fractionalization (Alesina et al. 2003) substitutes for ethnic fractionalization in our analysis of illiteracy. Because there is no standard benchmark model for any of these dependent variables we conduct two tests for each dependent variable. The first is a “full” model, including all variables discussed above. The second is a “reduced-form” model, omitting most controls that do not pass the threshold of statistical significance (p<0.10 in two-tailed tests), in the expected direction. We retain the geography-weighted control, the time-trend (for TSCS models), democracy stock, and GDP per capita variables in all models, regardless of statistical significance, because of our strong expectation that these variables capture important and otherwise unobserved effects. An appendix lists descriptive statistics as well as details about sources and coding procedures for all variables. RESULTS Table 1 presents the results of a series of tests of the empirical relationship between Parliamentarism and various governance outcomes. In each case, the model fit is quite good, with F values significant at better than the 0.0001 level and R2/pseudo R2 ranging from 0.47 to 0.90. Parliamentarism appears to be associated with some aspects of political development, 22 though the relationship is not robust across all dependent variables and specifications. Model 4 shows Parliamentarism to be correlated with the World Bank measure of corruption control, but only in the reduced-form model. Parliamentarism is strongly related to bureaucratic quality in both models (5 and 6), and with rule of law in the reduced form model (12). Results for the PRS measure of corruption control, government effectiveness and political stability are inconclusive. On balance, Parliamentarism appears to do no harm to political development (as measured here), and to be associated positively with certain aspects. Clearer patterns emerge in the area of economic development. Results suggest that Parliamentarism is associated with better telecommunications infrastructure, lower import duties, greater trade openness, better investment ratings, and higher levels of per capita GDP, across both full and reduced-form models. These results are robust in both the full and reduced form models, and across a wide variety of alternative specifications, most of which we do not report here. In two out of three cases, Parliamentarism is associated with higher levels of human development. All else being equal, longtime parliamentary systems have lower levels of infant mortality, and longer life expectancy. Again, results are robust across various specifications, and highly significant. No clear patterns emerge in the case of illiteracy. Table 2 summarizes findings for all regression tests reported in Table 1. It appears that parliamentary systems hold distinct advantages over presidential ones across a wide range of indicators of political, economic and human development. In no case are parliamentary systems associated with bad governance, and in most cases they are strongly associated with good governance. The results are particularly striking in economic and human development, where in every case but one a strong, positive relationship between Parliamentarism and good governance 23 can be seen. In addition to considering the statistical strength of the relationship between Parliamentarism and good governance, it is worth considering the substantive import of this relationship. A hypothetical scenario is instructive here. Employing the coefficients for Parliamentarism from the full models in Table 1 (and keeping all control variables constant), we find that fifty years of fully parliamentary rule (1951-2000, in this example) are associated with an improvement of 0.15 points on the seven-point scale of bureaucratic quality, forty-five percent more telephone mainlines, a five percent reduction in import duties as a percentage of imports, thirty percent more trade openness, more than six points higher investment rating (on a scale of 100), a thirty percent higher per capita income, a nearly twenty-three percent lower infant mortality rate, and more than a two percent greater life expectancy. Thus, while the substantive impact of Parliamentarism appears quite modest in some areas (e.g., bureaucratic quality and investment rating), it is quite large in others (e.g., telephone mainlines, trade and GDP per capita). Perhaps most importantly—for we do not wish to invest too much importance in imprecise point estimates—Parliamentarism shows a strong relationship with a wide variety of different outcomes, suggesting that we need to consider its overall cumulative effects across a quite broad spectrum of governance issues. ROBUSTNESS Several robustness issues arise. Consider, first, the construction of the main independent variable of interest, Parliamentarism. In our coding we recognize three levels: full presidentialism, semi-presidentialism, and full parliamentarism. This imposes an interval measurement onto a set of ordinal distinctions, an imposition that could result in significant measurement error. However, it is important to keep in mind that a) there are relatively few 24 semi-presidential systems, and b) they occupy the “in between” position in this trichotomy. Under the circumstances, it is unlikely that another method of coding would bring results significantly different from those reported here. Second, many of the countries with the longest history of parliamentary rule are located in Western Europe. These countries are also some of the best governed in the world. To control for this possibly confounding factor, we ran tests (not reported) including a dummy variable for Western Europe. Results for Parliamentarism remained quite robust, while the Western Europe variable itself performed inconsistently. Numerous additional specification tests were conducted, though we have no space to report them in Table 1. To our benchmark equations, we added additional control variables, such as a) the number of years each country has enjoyed national sovereignty (logged), b) decade dummies (to further control for time effects), and c) alternative measures of fractionalization (ethnic, linguistic, religious). None of these alternate specifications disturbed the results for Parliamentarism appreciably. We also experimented with the exclusion of various controls in the benchmark model. Again, we the results are reassuring. The relationship between Parliamentarism and good governance does not appear to be conditional on the inclusion of particular control variables in the general model. A fourth issue concerns the inclusion of observations during years when a country is only marginally democratic. Recall that we adopt a fairly low threshold for democracy (Polity2>0, on the Polity IV index, which stretches from -10 to +10). Our operating assumption is that a minimal level of multi-party competition is sufficient to set in motion whatever causal effects parliamentarism might have on the quality of governance. To the extent that the inclusion of less democratic cases weakens the potential impact of democratic institutions like Parliamentarism on 25 governance, this liberal inclusion criterion should bias the results against our hypotheses. However, it is possible that the effects of parliamentarism in a marginally democratic society are different than the effects in a fully democratic society. Reassuringly, additional tests that set the bar of democracy higher—at 5 on the 21-point Polity2 scale—yield little substantive difference in the results reported here. Note that observations in all regressions consist of country-years, not of country-cases. This means, perforce, that greater weight is carried in the analysis by countries that have been democratic over a longer stretch of time; the United States is a more influential case than recently democratized Namibia. As it happens, long-term democracies also tend to have stronger democracies; thus, the results, as reported, reflect a sample weighted towards countries that are strongly democratic. A fifth issue concerns possible interaction effects. It could be, for example, that parliamentary (and presidential) systems work differently in polities with different electoral system types (majoritarian, PR, preferential-vote, and so forth) or across unitary and federal systems. This may be probed empirically through interaction terms. While the empirical field of possible interaction terms is virtually limitless, we did not discover significant interaction effects with the main variable of interest, Parliamentarism. A final issue concerns possible selection effects. The assignment of the “treatment” in observational research is not random. This is problematic if the process of selecting political institutions is correlated with the outcomes under investigation (good governance, broadly considered). Perhaps, that is, parliamentary systems are more likely to be adopted where prospects for good governance are more propitious. In this case, our key variable will serve as a proxy for other, un-measured factors. In order to neutralize this potential identification problem we employed a series of instruments for Parliamentarism in two-stage least squares estimations. 26 Results so obtained are largely robust, thus providing at least some assurance that the effects reported here are not simply the product of a non-random choice of political institutions. Even so, we do not have a great deal of confidence in the two-stage models. All of the instruments that we could construct violate at least one of the assumptions of instrumental-variable analysis (Reiss 2003): they either correlate poorly with Parliamentarism or they correlate with the error term (i.e., they are probable causes of good/bad governance in the contemporary period). Thus, though the instrumental-variable results are supportive, we do not report them here. In any case, we think it unlikely that the choice of constitutional institutions reflects (rather than shapes) a country’s future prospects for good (bad) governance. To be sure, whether a country becomes parliamentary or presidential (or somewhere in between) depends partly on a country’s colonial heritage, its size and demographics, and on patterns of government that obtain in a regional or historical context. However, these exogenous influences are relatively easy to model and appear as controls in all our regression tests. Other factors influencing constitutional choice regarding the separation of powers do not seem to accord with a country’s proclivity to good or bad governance and may therefore be regarded as stochastic. One must also consider the fact that constitution-makers generally have notoriously short time-horizons. They are usually interested in installing a system that will benefit them personally, their parties, or their constituencies. In this respect, the type of constitution a country arrives at is the product of a highly contingent political battle, with little or no bearing on a country’s long-term governance potential. We think it reasonable to regard a country’s choices among constitutional institutions as a largely stochastic phenomenon with respect to the outcomes of interest in this study: long-term patterns of good or bad governance. One type of selection effect deserves special mention. It is sometimes argued that 27 democratic polities with presidential institutions are more liable to regime failure (e.g., Stepan & Skach 1993). If true (the point has been strenuously debated) then the sample may be skewed toward parliamentary systems, since we consider only democratic country-years in our empirical tests. However, this selection effect concerns independent variables (democracy and parliamentarism), not dependent variables; as such, it should not introduce bias into our estimates. Indeed, insofar as descent into autocracy may be considered a bad governance outcome, our exclusion of such country-cases from our analysis constitutes a conservative bias in the sample. We might expect stronger results if we did not drop these cases. CONCLUSION In an age of democratic transition, few questions have greater practical import than that of constitutional engineering. Yet, relatively little scholarly attention has been devoted to the impact of different types of constitutional design on policies and policy outcomes within democratic societies. The evidence presented here suggests that, to the extent that the nature of the executive makes a difference, parliamentary systems offer significant advantages over presidential systems. In no case examined here does parliamentary rule seem to detract from good governance. In most policy areas, particularly in the areas of economic and human development, parliamentary systems are associated with superior governance. It remains to explore the specific causal mechanisms by which the structure of the executive might influence various policy outcomes. To this point, we have merely reviewed some of the possibilities, as suggested by the voluminous literature on these questions. Plausible reasons why parliamentarism might lead to better governance include: a) stronger political parties, b) corporatist interest organization, c) tighter principal-agent relationships within the various arms of the bureaucracy, d) centralized (national-level) electoral accountability, e) the 28 capacity for flexible policymaking, f) a more institutionalized political sphere, and g) decisive leadership. The first section of the paper reviews each of these mechanisms. Is it possible to arrive at a relatively concise and coherent general theory that ties together and explains the findings of this (otherwise largely empirical) endeavor? To this end, we put it forth, provisionally and broadly, that what distinguishes parliamentarism from presidentialism, and what makes the former a more reliable vehicle for good public policy, is its capacity to function as a coordination device. The state is often conceptualized as a solution to the multiple coordination problems that emanate from society. There are many variations on this theme—joint-decision traps, shirking, under providing of public goods, overgrazing, the tragedy of the commons, common pool problems, collective action problems, free rider problems, prisoner’s dilemmas, transaction-cost dilemmas, and so forth. These are different ways of pointing out a central problem: quite often, when individuals or groups pursue their own agendas the result is not what society as a whole would prefer. Coordination problems thus involve a conflict between the part and the whole, between individual or group rationality and collective rationality. Arguably, parliamentarism offers better tools for resolving these sorts of difficulties than presidentialism (within the framework of democracy). This is because parliamentarism integrates a diversity of views, while providing greater incentives for actors to reach agreement. Consider a stylized comparison of decisionmaking processes within the two systems. Both feature a similar set of players: a legislature with majority and minority parties, committees and separate leadership hierarchies, a cabinet, a chief executive (PM or president), and various government agencies. Yet, the interaction of these players, and the role that each institution assumes, tend to be quite different. In parliamentary systems, debate occurs in a highly 29 institutionalized fashion—within parties, within committees, within leadership groups, across parties, within the cabinet, and between cabinet-level ministers and high-level civil service appointees within the bureaucracy. In presidential systems, by contrast, most of these units have greater independence, and those without independence (such as the cabinet) have very little power. The players that matter have the capacity, and often the incentive, to say no, or to insist upon side payments in exchange for support. While successful coordination can occur in this highly fragmented institutional sphere, resulting agreements may impose higher transaction costs than one would anticipate within a parliamentary system, where the incentives of the key actors are generally to reach agreement. This, in turn, is a product of how political careers and electoral incentives align within these various constitutional systems. The same sort of contrast operates at “mass” levels. In a parliamentary system, voters are encouraged to align their perspectives (interests, identity, ideals) with one of the national parties because these are the political institutions that matter. In a presidential system, voters may consider not only a candidate’s party membership but also her fealty to the district, along with various personal characteristics and issue-specific positions that may or may not be consistent with her party’s national strategy. Of course, we realize that the actual workings of parliamentarism/presidentialism within a polity depend upon many additional factors—sociological, economic, political, and historical—exogenous to the theory. Parliamentarism within a proportional representation electoral system works differently from parliamentarism in a first-past-the-post electoral system, for example. However, these additional considerations can for the most part be subsumed under the ceteris paribus caveat that accompanies all causal reasoning. Parliamentarism should be more successful in coordinating diverse views and interests than presidentialism, all other things 30 being equal. This means that the persistent institutional conflicts that characterize political life in all democracies—e.g., between legislature and executive, between backbenchers and party leaders, among parties, among diverse agencies, between national and subnational governments, among subnational governments, and among diverse constituencies—may be easier to solve in a parliamentary system than in a presidential system. This, in turn, should help to account for the higher quality of governance observed in parliamentary systems across many policy areas. 31 Table 1: Estimation Results Corruption Control (PRS) Political development Model Estimator Parliamentarism (historical) Geography-weighted DV Trend Democracy (historical sum, ln) GDP per capita, ln (1960) Africa Asia LAC Socialism English legal origin Latitude, ln Ethnic fractionalization Population, ln (1960) Distance from financial center Oil production, per capita Diamond production, per capita Protestant Constant Observations Countries Sample period R2/Pseudo R2 Prob > F Corruption Control (WB) Bureaucratic Quality (PRS) Government Effectiveness (WB) Political Stability (WB) Rule of Law (WB) 1 2 3 4 5 6 7 8 9 10 11 NW NW BE BE NW NW BE BE BE BE BE 12 BE 0.001 0.001 0.003 0.004** 0.002** 0.003*** 0.0002 0.002 0.0003 0.0003 0.003 0.005*** (0.002) (0.002) (0.002) (0.002) (0.001) (0.001) (0.002) (0.002) (0.003) (0.002) (0.002) (0.002) 0.002 0.011* 0.002 0.026* 0.027*** 0.028*** 0.008 0.034** 0.020 0.017 0.026 0.042*** (0.009) (0.006) (0.018) (0.015) (0.008) (0.005) (0.018) (0.015) (0.032) (0.029) (0.020) (0.016) -0.031*** -0.026*** 0.011** 0.010** (0.007) (0.007) (0.005) (0.005) 0.637*** 0.337** 1.083*** 0.888*** 0.598*** 0.507*** 0.994*** 0.876*** 0.671*** 0.587*** 0.869*** 0.680*** (0.184) (0.149) (0.189) (0.153) (0.121) (0.098) (0.190) (0.150) (0.248) (0.191) (0.190) (0.151) 0.480*** 0.459*** 0.158*** 0.189*** 0.475*** 0.509*** 0.201*** 0.221*** 0.155*** 0.149*** 0.207*** 0.184*** (0.044) (0.032) (0.042) (0.037) (0.036) (0.027) (0.043) (0.037) (0.059) (0.048) (0.042) (0.037) -0.125 -0.175 -0.366*** -0.167 0.286 0.027 (0.169) (0.175) (0.134) 0.263* 0.010 0.407*** 0.630*** (0.177) (0.254) (0.174) 0.054 0.106 (0.147) (0.163) (0.114) 0.285* (0.084) (0.164) (0.231) (0.163) -0.902*** -0.814*** -0.197 -0.174* -0.510*** -0.479*** -0.245* -0.225** 0.020 -0.075 (0.120) (0.098) (0.134) (0.100) (0.100) (0.078) (0.134) (0.100) (0.206) (0.132) 0.484*** 0.131 0.260** 0.047 0.159 0.177 (0.177) (0.169) (0.120) (0.170) (0.223) (0.169) -0.391*** -0.141 0.436*** 0.393*** -0.080 -0.129 -0.037 (0.085) (0.107) (0.063) (0.059) (0.109) (0.145) -0.068 0.297*** 0.280*** -0.095* 0.267*** 0.275*** 0.339*** 0.416*** 0.249*** 0.281*** (0.049) (0.073) (0.059) (0.052) (0.074) (0.060) (0.110) (0.090) (0.076) (0.060) 0.130 -0.012 -0.057 0.159 -0.396 (0.180) (0.107) 0.036 (0.205) (0.126) (0.207) (0.307) -0.079*** -0.053*** -0.018 0.071*** 0.016 -0.113*** -0.112*** (0.204) 0.003 (0.020) (0.019) (0.024) (0.015) (0.024) (0.037) (0.032) (0.024) 0.033* -0.009 0.003 -0.021 -0.013 -0.030 (0.018) (0.022) (0.012) (0.022) (0.030) (0.022) -0.776*** -0.607** -0.397 -0.470** -0.298* -0.305 0.058 -0.291 (0.280) (0.291) (0.609) (0.239) (0.175) (0.614) (0.765) (0.609) 0.038* 0.056* 0.070*** 0.059* 0.026 0.075** (0.020) (0.033) (0.018) (0.034) (0.044) (0.033) 0.009*** 0.008*** 0.007*** 0.005*** 0.001 0.003 0.006*** 0.004** (0.001) (0.001) (0.002) (0.002) (0.001) (0.002) (0.002) (0.002) -0.393 1.160 -7.759*** -7.040*** -7.887*** -6.213*** -7.900*** -7.042*** -3.192* -2.621** -7.070*** -5.620*** (1.410) (1.315) (1.296) (0.941) (0.919) (0.786) (1.302) (0.916) (1.736) (1.268) (1.300) (0.925) 1218 97 1244 99 311 118 323 123 1218 97 1251 100 328 118 341 123 294 106 297 107 321 119 333 124 1984-2001 0.64 0.0000 0.64 0.0000 1996-2000 0.84 0.0000 0.82 0.0000 1984-2001 0.73 0.0000 0.72 0.0000 32 1996-2000 0.80 0.0000 0.78 0.0000 1996-2000 0.62 0.0000 0.60 0.0000 1996-2000 0.82 0.0000 0.80 0.0000 Telephone Mainlines (ln) (WDI) Economic development Model Estimator Parliamentarism (historical) Geography-weighted DV Trend Democracy (historical sum, ln) GDP per capita, ln (1960) Africa Asia LAC Socialism English legal origin Latitude, ln Ethnic fractionalization Population, ln (1960) Distance from financial center Oil production, per capita Diamond production, per capita Constant Observations Countries Sample period R2/Pseudo R2 Prob > F 13 NW Import Duties (WDI) Trade openness (ln) (WDI) Investment Rating (Euromoney) GDP per capita (ln) (WDI) 14 15 16 17 18 19 20 21 NW NW NW NW NW NW NW NW 0.006*** 0.003*** (0.001) (0.001) -0.001* -0.001** (0.001) (0.0005) 0.049*** 0.051*** (0.003) (0.002) 0.174 0.212* (0.136) (0.128) 0.709*** 0.765*** (0.068) (0.058) -0.799*** -0.780*** (0.140) (0.102) -0.371** (0.146) 0.294*** (0.076) 0.499*** 0.406*** (0.148) (0.147) 0.038 (0.060) 0.270*** 0.251*** (0.053) (0.050) -0.124 (0.115) -0.060*** -0.100*** (0.012) (0.011) -0.052*** -0.052*** (0.010) (0.011) -0.722*** -0.446** (0.142) (0.184) 0.062*** (0.015) -5.128*** -5.338*** (0.769) (0.730) 2129 2171 126 128 1965-2001 0.79 0.78 0.0000 0.0000 -0.067*** -0.058*** (0.010) (0.010) -0.030 -0.009 (0.018) (0.014) -0.093*** -0.128*** (0.032) (0.029) 3.785*** 6.288*** (1.303) (0.786) -2.939*** -3.557*** (0.535) (0.367) 2.980* (1.642) 2.635** (1.098) 0.052 (0.900) -3.909*** (1.177) 1.133** (0.490) 0.502 (0.562) -0.901 (1.200) -0.628*** (0.194) 0.487*** 0.729*** (0.097) (0.095) 0.594 (1.465) -0.237 (0.176) 24.577*** 3.310 (8.399) (5.797) 1411 1436 97 98 1970-2000 0.50 0.47 0.0000 0.0000 0.004*** 0.004*** (0.001) (0.001) -0.0002 0.002* (0.001) (0.001) 0.008*** 0.007*** (0.001) (0.001) -0.073 -0.203*** (0.050) (0.041) -0.020 -0.043*** (0.015) (0.010) -0.144*** (0.051) 0.027 (0.051) -0.248*** -0.175*** (0.033) (0.025) 0.286*** (0.052) 0.015 (0.026) -0.054*** (0.019) 0.372*** (0.049) -0.246*** -0.225*** (0.006) (0.006) -0.045*** -0.047*** (0.005) (0.004) -0.0003 (0.086) 0.018*** (0.007) 7.753*** 8.748*** (0.375) (0.317) 2521 2611 126 132 1961-2001 0.68 0.63 0.0000 0.0000 33 0.085*** 0.089*** (0.020) (0.020) 0.008* 0.009* (0.005) (0.005) 0.296*** 0.276*** (0.085) (0.086) 12.786*** 12.354*** (2.347) (2.268) 8.598*** 8.893*** (0.577) (0.544) -4.034** (1.689) 7.420*** 8.744*** (1.805) (1.752) -11.035*** -9.470*** (1.442) (1.197) -9.997*** -9.317*** (2.170) (2.097) -0.417 (0.980) 1.900** 1.832** (0.852) (0.854) -2.307 -4.060** (1.995) (1.926) 2.841*** 2.919*** (0.236) (0.225) -0.900*** -1.050*** (0.233) (0.228) -0.684 (2.392) 1.427*** 1.295*** (0.211) (0.206) -163.172*** -162.295*** (17.637) (17.709) 1544 1544 122 122 1982-2001 0.81 0.81 0.0000 0.0000 22 NW 0.004*** 0.003*** (0.001) (0.001) 0.037*** 0.036*** (0.014) (0.014) 0.014*** 0.014*** (0.001) (0.001) 0.348*** 0.317*** (0.102) (0.095) 0.767*** 0.767*** (0.065) (0.051) -0.186* -0.273*** (0.096) (0.067) 0.100 (0.097) 0.119** (0.049) -0.281*** -0.341*** (0.098) (0.104) -0.048 (0.038) 0.192*** 0.178*** (0.042) (0.045) -0.291*** -0.279*** (0.071) (0.074) -0.022*** -0.020*** (0.007) (0.007) -0.049*** -0.048*** (0.005) (0.006) 0.334** 0.402** (0.150) (0.162) 0.130*** 0.130*** (0.007) (0.007) -0.830* -0.650 (0.445) (0.426) 2522 2522 124 124 1961-2000 0.90 0.90 0.0000 0.0000 IMR (ln) (WDI) Human Development Model Estimator Parliamentarism (historical) Geography-weighted DV Trend Democracy (historical sum, ln) GDP per capita, ln) (1960) Africa Asia Life Expectancy (ln) (WDI) 23 24 25 26 27 28 NW NW NW NW NW NW -0.003*** -0.004*** 0.0003*** 0.0003*** -0.001 -0.001 (0.001) (0.001) (0.0001) (0.0001) (0.001) (0.001) 0.002* 0.002* 0.00002 -0.0002* 0.055*** 0.057*** (0.001) (0.001) (0.0001) (0.0001) (0.008) (0.008) -0.031*** -0.030*** 0.003*** 0.003*** -0.020*** -0.021*** (0.001) (0.001) (0.0002) (0.0002) (0.002) (0.002) -0.522*** -0.508*** 0.027** 0.019** -0.564*** -0.503*** (0.066) (0.061) (0.011) (0.008) (0.119) (0.116) -0.295*** -0.291*** 0.046*** 0.042*** -0.746*** -0.769*** (0.036) (0.024) (0.006) (0.005) (0.031) (0.025) 0.394*** 0.389*** -0.191*** -0.206*** 0.218*** 0.257*** (0.070) (0.055) (0.014) (0.013) (0.084) (0.068) -0.090 (0.086) LAC Socialism English legal origin Latitude, ln Population, ln (1960) Distance from financial center Oil production, per capita Diamond production, per capita Muslim Linguistic fractionalization Constant Observations Countries Sample period R2/Pseudo R2 Prob > F 0.023** -0.131 (0.012) (0.097) 0.197*** 0.175*** 0.019** 0.517*** 0.589*** (0.051) (0.043) (0.008) (0.094) (0.077) -0.448*** -0.501*** 0.019 -2.043*** -2.019*** (0.078) (0.080) (0.012) (0.158) (0.149) 0.021 -0.016*** -0.017*** -0.311*** -0.315*** (0.027) (0.006) (0.006) (0.050) (0.050) -0.040 0.016*** 0.014*** -0.073* (0.044) (0.034) Ethnic fractionalization Illiteracy (ln) (WDI) (0.005) (0.005) 0.501*** 0.521*** -0.066*** -0.060*** (0.059) (0.049) (0.011) (0.012) 0.031*** 0.017** -0.007*** -0.006*** 0.020* (0.008) (0.007) (0.001) (0.001) (0.011) 0.042*** 0.044*** 0.001 0.060*** 0.062*** (0.005) (0.006) (0.001) (0.008) (0.008) -0.313*** -0.001 -0.227 (0.108) (0.024) (0.276) 0.002 -0.003 0.011* (0.010) (0.005) (0.006) 0.005*** 0.005*** -0.001*** -0.001*** 0.009*** 0.009*** (0.001) (0.001) (0.0001) (0.0001) (0.001) (0.001) 0.715*** 0.709*** (0.088) (0.087) 10.851*** 10.977*** 3.543*** 3.624*** 11.865*** 11.971*** (0.402) (0.388) (0.065) (0.046) (0.852) (0.767) 2633 126 0.82 0.0000 2682 129 1961-2001 0.81 0.0000 2634 125 0.75 0.0000 NW=Newey-West 2652 127 1961-2001 0.75 0.0000 2401 108 0.85 0.0000 2438 110 1961-2001 0.85 0.0000 standard errors, including correction for first-order autocorrelation. BE=Between-effects OLS regression. * p<10%;** p<5%;*** p<1% All samples limited to country-years that are minimally democratic (Polity2>0) 34 Table 2: Summary Results OUTCOME SPECIFICATION MODELS (TABLE 1) POLITICAL DEVELOPMENT Corruption control (PRS) Corruption control (WB) Bureaucratic quality (PRS) Government effectiveness (WB) Political Stability (WB) Rule of Law (WB) ECONOMIC DEVELOPMENT Telephone mainlines (WDI) Import Duties (WDI) Trade openness (WDI) Investment rating (Euromoney) GDPpc HUMAN D EVELOPMENT IMR (WDI) Life expectancy (WDI) Illiteracy Full Reduced ++ ++ +++ +++ 1, 2 3, 4 5, 6 7, 8 9, 10 11, 12 +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ 13, 14 15, 16 17, 18 19, 20 21, 22 +++ +++ +++ +++ 23, 24 25, 26 27, 28 ++ and +++ indicate statistical significance at the 0.05 and 0.01 levels, respectively, in the expected direction (that is, models in Table 1 that show Parliamentarism to be positively correlated with political, economic, or human development) 35