A review of the political economy of governance: From property rights to voice Philip Keefer Development Research Group World Bank pkeefer@worldbank.org Paper for project on "The Consequences of Political Institutions in Democracy" Coordinated by Herbert Kitschelt, Duke University Abstract: This paper reviews progress made in understanding the effects of different dimensions of governance on economic development, and the sources of "good governance." The term governance has been used to embrace concepts that are heterogeneous both with respect to their effects on economic development and their genesis. Future progress in developing policy responses to "bad governance" will depend on separately examining these heterogeneous elements ­ the security of property rights, the quality of bureaucratic performance, corruption, voice and accountability. Future progress will also depend on explicitly linking problems of governance to the overarching political environment and the incentives of governments to correct those problems. World Bank Policy Research Working Paper 3315, May 2004 The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working Papers are available online at http://econ.worldbank.org. A review of the political economy of governance: From property rights to voice A growing body of evidence points to governance failures as a root cause of slow and inequitable economic growth and as a defining characteristic of most poor countries. These findings justify placing governance high on any research agenda aimed at better understanding the political economy of economic development. Already, research into governance and development has had a notable impact: some dimensions of governance now sit at the center of academic and policy discussions of economic development. This paper reviews the known effects of governance on development, the interrelationship among the different dimensions of governance, and the origins of "good" governance. The review highlights where important questions remain to be answered, particularly with respect to the origins of good governance. Although this paper is a critical review of the governance literature, it turns out that most of that literature does not use the term "governance." Moreover, there is no agreed definition of governance that would provide a convenient device for organizing the literature. Finally, there are few research efforts that set out to analyze all dimensions or even most dimensions of governance jointly. For various, sometimes necessarily arbitrary reasons that are explained below, the focus here is therefore on the literature that links economic development to secure property rights, voice and accountability, or the performance of the bureaucracy. Each of these seems to be at the core of all definitions of governance. The basic conclusions of this review are threefold. First, further research on governance-related issues remains a high priority, but progress is likely to be fastest and 2 most convincing when future work addresses the components of governance rather than aggregated concepts of governance. In many cases, the components of governance do not even bear the same causal relationship to development, nor is one component necessarily a good proxy for other components. The security of property rights, for example, can be considered a proximate contributor to economic development, in much the same way as macroeconomic or social policy. Voice and accountability, however, matter indirectly, through their influence on government decision making or the security of property rights. The review spells out the differences across governance indicators and their links to development. The second conclusion of this review is that evidence and theory better support the influence of some components of governance on development than others. The security of property rights and the credibility of governments emerge as the components with the best documented and strongest influence on economic development. Causality problems cloud estimates of the influence of bureaucratic (or "state") capacity and corruption on development. The most critical of these causality problems results from the omission from most analyses of political variables that are likely to influence both bureaucratic efficiency and integrity and development outcomes. Finally, analyses of voice and accountability, or "democracy", while the subject of substantial attention, have suffered from a lack of theoretical and empirical precision that clouds interpretation. Third, future research that deepens our understanding of the determinants of good governance is likely to have the greatest payoff. Although progress has been made in identifying the political and social conditions that lead to more secure property rights, greater voice and accountability, or more efficient and honest bureaucracy, much remains 3 unknown or puzzling. In addition, the governance literature has so far been isolated from much of the progress that has been made along these lines in the broader political economy literature. A growing literature outside the realm usually defined as governance describes the effects of voter information and political institutions on political incentives to seek rents; the tools and results of this literature have yet to be integrated into work on governance. What is governance? Whether in policy or academic settings, governance is among the more elastic concepts in the social science and development lexicons. Definitions tend to encompass one or both of the following: the extent to which governments are responsive to citizens and provide them with certain core services, such as secure property rights and, more generally, the rule of law; and the extent to which the institutions and processes of government give government decision makers an incentive to be responsive to citizens. Though similar, in fact the first are "outcomes" while the second are "causal" or more fundamental concepts. Corruption and bureaucratic quality are more direct indicators of lack of responsiveness, and only indirect indicators of government incentives; measures of democracy or voice and accountability, in contrast, directly capture the second more than the first. The Institute on Governance, in Canada, defines governance squarely in the second category, as comprising "the traditions, institutions and processes that determine how power is exercised, how citizens are given a voice, and how decisions are made on issues of public concern."1 The journal Economics of Governance essentially views 1http://www.iog.ca/about_us.asp?strTextSite=false. 4 governance as synonymous with governing, and encourages submissions that deal with all manner of problems that emerge in the way groups of individuals govern themselves in the public or private sectors. Its statement of aims and objectives, however, evades the tricky issue of how to define governance. More common are definitions and realms of activity that straddle the outcome and institutional sides of governance. Governance for the OECD relates to "institutions, policy making and participation of civil society."2 The Ford Foundation views governance as the extent to which government institutions are "transparent, accountable, responsible and guided by the rule of law and dedicated to reducing inequality."3 The US Agency for International Development focuses on outcomes in defining good governance: it is present when governments can "maintain social peace, guarantee law and order, promote or create conditions necessary for economic growth, and ensure a minimum level of social security." However, underlying good governance are transparent and accountable government institutions, and USAID work on governance focuses explicitly on strengthening democratic institutions.4 The United Nations Development Program refers to "democratic governance," but its description of its goals implies that governance is the responsiveness of state institutions and processes to the needs of ordinary citizens.5 Perhaps not surprisingly, since their charters discourage engagement with countries on more overtly political issues, the least political definitions of governance can be found at the World Bank and International Monetary Fund. The World Bank 2http://www.oecd.org/EN/about/0,,EN-about-333-9-no-no-no-0,00.html 3http://www.fordfound.org/ 4http://www.usaid.gov/democracy/gov.html 5http://www.undp.org/governance/ 5 introduces its work on governance and links it intrinsically to public sector reform, characterizing governance and public sector reform jointly as focused on "building efficient and accountable public sector institutions". Governance, then, is the extent to which public sector institutions are accountable and capable of sustaining development. The emphasis, however, is on the implementing or administrative agencies of government rather than the incentives of the political actors who sit atop them.6 Similarly, the International Monetary Fund views good governance as encompassing (but not necessarily exclusive to) the rule of law, the efficiency and accountability of the public sector, and corruption.7 The various definitions of governance are evidently expansive. In fact, if the study of governance extends to all questions related to how groups of people govern themselves, as might be inferred from many of these definitions, then there are few subjects in all of political science and political economy that do not fall within the governance domain. Taken to this extreme, one might reasonably doubt whether the study of governance is at all novel or of independent analytic interest. There are two responses to this. First, in one sense it does not really matter. As long as the literature labeled governance generates important findings about critical topics and does not ignore the broader research domain in which those topics may be nested, the label is irrelevant. Second, it is still possible to establish useful ­ if somewhat arbitrary ­ boundaries on governance as an independent sphere of inquiry. In both policy and academic circles the term has acquired popularity only recently and with two particular concerns. Governance is first associated with establishing causal relationships 6http://www1.worldbank.org/publicsector/ 7http://www.imf.org/external/pubs/ft/exrp/govern/govindex.htm 6 between economic development and particular performance characteristics of developing countries to which earlier literature paid less attention. These characteristics include, especially, such phenomena as the security of property rights, the performance of bureaucracies and the predictability and credibility of government decision making. Second, governance is associated with establishing causal relationships between the underlying institutional characteristics of governments (typically, democracy or voice and accountability) and either the performance of governments (e.g., with respect to the rule of law) or economic development, or both. The discussion below therefore takes these two lines of inquiry as comprising the study of governance. Although the literature does not always follow this division and often conflates the study of less and more fundamental notions of governance, in principle these are clearly two different and well-defined lines of inquiry that mark a literature that has grown in rigor and volume since the 1980s. The hazards of aggregation in governance research Policy and some academic discussions frequently refer to good governance, implicitly aggregating all of the underlying dimensions into a single concept. Kaufmann, Kraay and Zoído-Lobatón (2002) reflect the breadth of this concept in their efforts to devise systematic measures of governance. They take 194 measures variously used in the literature as measures of governance, from 17 sources, and divide them into six categories: voice and accountability, political stability, government effectiveness, regulatory quality, rule of law and control of corruption. Theirs is a reasonable breakdown of governance concepts, and there are few analyses of governance, implicit or 7 explicit, that do not fall into one or more of these categories. Aggregation of these concepts makes sense if all good things go together. In fact, they may not. Kaufmann and Kraay (forthcoming) highlight the difficulty of casually aggregating these governance components. They show that good performance on some of these dimensions does not imply good performance on others. Most countries in Latin America turn out to do better than other countries with similar per capita incomes with respect to voice and accountability, but worse with respect to government effectiveness, rule of law and corruption, and both better and worse with respect to political stability and regulatory quality. Recognizing this, in their substantive analysis of the effect of governance on growth they focus only on one of their indicators, the rule of law. Studies of specific countries reveal how divergence among indicators might emerge. Figure 1 records the assessment by one political risk service of corruption perceptions in Indonesia. It shows clearly that corruption perceptions rose with the advent of democratic elections on June 7, 1999 and the demise of the Suharto autocracy. Voice and accountability therefore seem to have moved inversely with corruption perceptions. The paradox here is acute: the Suharto regime was widely regarded as among the most corrupt in the world, and no observer doubts that the absolute value of bribes going to the government has fallen precipitously, despite worsening corruption perceptions. One explanation for this paradox is that corrupt transactions became less credible after Suharto's departure, suggesting that voice and accountability (at least as generated by new and imperfect democracies) may also diverge from government credibility. Under Suharto, businesses believed that if they paid a $1,000,000 bribe they would get a 8 high return on their investment because the underlying agreement was credible (or, in terms of the Kaufmann, Kraay and Zoído-Lobatón variables, either the rule of law or political stability were high). They could be confident that they would, in fact, receive the rents conferred by the monopoly or regulatory privilege provided in exchange for the bribe. The political uncertainty of the post-Suharto era lowered the credibility of these transactions. Consequently, even if the bribe-price of entry or regulatory privileges fell since the end of the Suharto regime, the effective value of the privileges may have fallen by even more. Although total corrupt payments may have fallen, the perceived damage of corruption might have risen. The Indonesia case demonstrates that two components of governance ­ corruption and the credibility of government ­ may also move in opposite directions, at least as they are measured. Figure 1: The Evolution of Corruption Perceptions, Indonesia, 1995 - 2001 4 3 2 1 0 1995 1996 1997 1998 1999 2000 2001 Note: Political Risk Services Corruption assessment, 0 ­ 6 scale, higher implies less corruption. One final example of the potential divergence among typical dimensions of governance relates to the credibility of government and bureaucratic quality. North and Weingast argue that credibility is a function of checks and balances among political decision makers. We can therefore imagine two governments, one exhibiting checks and 9 balances, and one not. Which will have an easier time ensuring a high level of performance by civil servants? Keefer and Stasavage (2002) argue that independent central banks are much more likely to succeed in generating low inflation outcomes in the presence of checks and balances. This seems to suggest that credibility and bureaucratic quality might go together. However, problems of special interest capture of regulatory agencies are well-documented in many other contexts. To the extent that checks and balances insulate captured regulators from political interference, we might expect poor regulatory/bureaucratic performance to be greater when political institutions are associated with more credibility. Contributors to Libecap (1996) make exactly this point and Wallsten (2002) has documented worse outcomes in electricity regulation when regulators are fully independent of political authorities (when they cannot be removed from office and they have an independent source of financing). An abridged intellectual history of governance The concerns motivating the governance literature have a distinct intellectual heritage. Broadly speaking, of course, analysis of the operation of bureaucracies and their honesty and efficiency dates back to Max Weber. The connection to economic development, a crucial component of the governance literature, is more recent. Similarly, though research on democracy and democratization goes back generations, it is only more recently that scholars have begun to ask systematic questions relating the incentives of government actors to outcomes in developing countries. In the 1950s, the common wisdom in policy and academic circles characterized the key obstacle to development as capital, so that economic development (or at least growth) depended on the transfer of capital to capital-starved poor countries. By the 10 1980s development thinking had shifted to concern with the economic policies of developing country governments, typically trade barriers, state ownership of industry, and loose management of the macro-economy. Experience and research then showed that reasonable economic policies were possibly a necessary but certainly not a sufficient condition for economic growth and development. This motivated a renewed look at other aspects of public sector performance. The work of North (1981, 1990) was particularly important in setting the intellectual stage for this new focus. He systematically linked country characteristics such as the security of property rights directly to the prosperity of nations. From this evolution, then, emerged one part of the governance research agenda: improved understanding of the effects of the non-policy characteristics of government performance on economic development. The research stemming from this strand of the governance literature brought to the forefront the contribution to development of such previously under-examined phenomena as the security of property rights, the rule of law, expropriation, bureaucratic quality, red tape, and the quality of government regulation ­ concepts falling squarely under the rubric of governance. While these ideas flowered in the academic literature, they were also given great policy impetus by work not primarily intended for the research community, particularly the controversial, though influential contribution of de Soto (1989). A second pillar of development wisdom in the 1950s ­ or, at least, an unexamined and implicit assumption ­ was that governments would, by and large, act in the general interests of society. Most explanations of underdevelopment ignored the incentives of the governments of developing countries and focused instead on exogenous or purely 11 economic constraints on development or, in the political sphere, the role of industrialized countries in perpetuating underdevelopment. This was a reasonable focus for the times, since the colonial legacy was both recent and bitter in many developing areas. Bates (1981) broke with this tradition by documenting how the political or economic self-interest of political actors could yield policies that were devastating for society at large, independent of external influence. He specifically showed that governments in the countries he considered sought the support of powerful interest groups and favored them at the expense of the unorganized or less threatening interest groups. This work inspired a new question: why did interest groups in the African countries that he examined have a more pernicious effect on outcomes than interest groups elsewhere ­ in Germany or the United States, for instance? Under what conditions, that is, do governments in some countries pursue socially beneficial policies while in other countries they do not? From these questions flow the second part of the governance research agenda: under what conditions do governments pursue the public interest and promote development in poor countries? This overly ­ even unreasonably ­ condensed intellectual history of governance excludes some concepts and literature that others reasonably include within the governance domain. In particular, the inadequacy of development explanations rooted in capital inflows and economic policy choices led many researchers to focus not on self- interested political leaders but rather on broader characteristics of the state or the bureaucracy. This research has introduced concepts such as the rent-seeking state (Buchanan, et al. 1980), embedded autonomy (Evans 1995), and state capacity using 12 analytical approaches that are useful to contrast with the more overtly political analyses spawned by Bates. The literature on rent-seeking generally takes for granted that governments use the coercive power of the state for self-aggrandizement. It does not take into account the limits that political competition might impose on these efforts. Arguments about embedded autonomy and state capacity take a more benign view of the state than the rent-seeking literature. They emphasize the role of the administrative apparatus of the state in promoting development. Like the rent-seeking literature, however, this literature does not specify the conditions under which politicians support or demand high performance from the bureaucracy. In both cases, therefore, this work abstracts from the political considerations that are at the core of arguments such as those of Bates (1991). The conclusion of Evans (1989), for example, suggests that politicians are but background or ancilliary actors in development: "The comparative evidence suggests that the efficacy of the developmental state depends on a meritocratic bureaucracy with a strong sense of corporate identity and a dense set of institutionalized links to private elites." This evidence could be persuasive if the politicians in the different countries confronted approximately the same electoral and institutional constraints on their actions. This may not be the case, however, since the conclusion rests on a comparison of Brazil, Japan and Zaire. Evans (1989) acknowledges some of these uncertainties, pointing out that despite evidence of the connection between development success and meritocratic bureaucracies working hand in hand with the private sector, "[t]he constraints that keep the inevitable clientelism and corruption from overwhelming the utility of ties to private capital are still not defined" (p. 583). 13 At the same time, however, analyses such as those of Evans and, in a different tradition, those contained in Libecap (1996), suggest that it may also be problematic to ignore the independent role for the bureaucracy in public policy formation. The argument made below, however, is that the analysis of political incentives without taking into account bureaucratic capacity may be a more tenable strategy than the reverse. Governance and economic development: Reviewing the theory The governance literature is largely empirical. The successes and failures on the empirical side of the literature, however, are related to the clarity and precision of its theoretical underpinnings. Clarity is of course inherently difficult to achieve with a concept as heterogeneous as governance. It helps to consider two sets of governance concepts separately, those related to government performance of its core functions and those related to the incentives of government officials to perform those functions. The former is the most successful strand of the governance literature, making the argument that secure property rights, predictable and credible government, and honest and efficient bureaucracy have a significant impact on economic development. The second strand of the governance literature concerns the incentives of government actors to act in the public interest. This strand has confronted greater challenges and is the most promising and important for future research. Within the first strand, the theoretical arguments are not equally strong or unambiguous. Least disputed and most clear is the theory linking insecure property rights to slow growth. Similarly, credible government is generally and unambiguously regarded as good for public policy. The frequent confusion of credibility with predictability undermines clarity, however, since the latter is not strongly related to 14 development and the quality of public policy. However, as the foregoing brief references to embedded autonomy and state capacity suggest might be the case, the influence of an honest and efficient bureaucracy on economic development is most ambiguous of all. On the one hand, this influence depends, as a matter of logic, on what the bureaucracies are doing; on the other hand, what bureaucracies do is generally controlled by politicians. Each of these are considered separately in the sections that follow. The second strand of the governance literature focuses on voice, accountability and democracy and their effects on growth and development. It pays less attention to specific implications of different electoral and political institutions within democracies and to the dynamics of electoral competition, all of which vary across countries. In this sense, it is somewhat divorced from a sophisticated and rapidly growing literature exploring the political and electoral determinants of politician incentives to act in the public interest. Nevertheless, the introduction of notions of voice, accountability and democracy have been critical in pointing development research towards the question, "Under what conditions do governments have the incentive to ensure the rule of law, secure property rights, and a well-functioning bureaucracy, and, more generally to serve the public interest?" The security of property rights and economic growth The theoretical case for secure property rights is simple: growth is a prerequisite for economic development (even if not a sufficient condition for it). Growth is not possible without investment. However, investors do not invest when they fear confiscation of their assets by government. North (1981) makes this point repeatedly, and a plethora of formal models explicitly chart the path from property rights insecurity 15 to slow growth.8 In its bare essentials, few propositions are less controversial than this argument. Still, there are objections to the theoretical case linking the security of property rights to growth. One relates to definitional confusion; the other to the effects of institutional influences that might explain both the security of property rights and economic growth. Definitional confusion emerges because two important notions of property rights are often confused, the allocation of property rights and the security of property rights. The governance literature deals only with the latter: do property owners have protection from the arbitrary confiscation of their assets (either through expropriation, tax law changes, or through interpretations of existing regulation)?9 Discussions of central planning, the transition from communism, and the costs and benefits of privatization address the other notion of property rights, their allocation. Przeworski and Limongi (1993) provide one example of the ease with which the two notions can be conflated. They point to arguments that democracy may render property rights less secure because the introduction of democracy creates opportunities for the poor to redistribute incomes away from the rich.10 However, it is not democracy, per se, that creates insecurity, but the transition to democracy. In equilibrium, or once democracy is established, there is no reason to expect the distribution of property rights to change further. Property rights, 8Keefer and Knack (2002b), for example, show that whether property rights insecurity is the product of a general level of insecurity in a society that makes it costly for governments to protect property rights, or as the product of particularly short horizons of government leaders that makes it more likely for them to prefer expropriation over growth, property rights insecurity slows growth. 9It should not go without saying that other, non-economic rights tend also to be insecure in these same countries. 10They point to Marx' argument that capitalism and democracy are inherently incompatible, since democracy enables the poor to tax the rich, giving the rich an incentive to buy military intervention to overturn the democracy, an argument formalized by Acemoglu and Robinson (2001). 16 after the initial redistribution, can therefore still be secure.11 Similarly, Rodrik, et al. (2002) argue that the success of countries such as China create difficulty for the thesis that property rights matter for growth. Again, they are referring to the allocation of property rights rather than the security of the allocation. As long as economic actors are confident that the rules will not change arbitrarily, even if the rules imply high tax rates, their rights are secure; it is when the rules are subject to arbitrary and significant alteration ­ regardless of whether they are initially favorable or not to the economic actors ­ that rights are insecure. There is ample evidence of this. Countries that have the highest scores on the security of property right measures, such as Australia, Germany, New Zealand and the United States, collect taxes ranging from 18 percent of national income (the US) to 36 percent (Sweden). Tax collection in countries rated as posing high risks of contract repudiation or expropriation are often in the teens and rarely above 25 percent of national income.12 The second objection that arises to the property rights arguments in the governance literature relates to causality. As Weyland (in this volume) carefully describes, arguments linking property rights to growth are sometimes asserted to be vulnerable to both of the usual causality flaws: either the association of secure property rights and economic growth is a reflection of reverse causality, from the second to the first, or, more difficult and interesting, the security of property rights and economic growth are both determined simultaneously by some third and omitted factor. 11 Robinson and Acemoglu (2001) develop a model in which, when the cost of coups by the rich against poor democracies and wealth inequality are in precisely the right combination, democracy is chronically unstable. The insights they develop from this model are significant, but the empirical relevance of this particular case has not yet been established. 12 There is little evidence that redistribution ­ or more broadly, the allocation of property rights ­ is systematically associated with low growth. See, for example, Knack and Keefer (1997). 17 An important contribution of Acemoglu, Johnson and Robinson (2001) to scholarship on property rights and growth is an ingenious econometric instrument that they discovered for the security of property rights (settler mortality rates in the 18th and 19th centuries). They use this variable as an econometric instrument for the property rights variable used in Knack and Keefer (1995), though they refer not to the security of property rights but to "institutions". They emphasize that their contribution is large because institutions are notoriously endogenous: "It is quite likely that rich economies choose or can afford better institutions. Perhaps more important, economies that are different for a variety of reasons will differ both in their institutions and in their income per capita" (p. 1369). Ironically, the theoretical (though not empirical) reasons to be concerned about whether secure property rights "cause" economic growth are not particularly convincing.13 If property rights are expensive to secure, for example, one might argue that rich countries are more likely than poor countries to secure them; fast-growing countries would therefore be more likely to improve property rights. However, to the extent that the security of property rights relates to the absence of expropriatory decisions by the state, there seem to be few financial costs to protecting property rights and little reason to argue that secure property rights should be sensitive to wealth. Even elaborate judicial systems are not particularly expensive. Alternatively, it might be that citizens of rich countries prefer secure property rights more than citizens of poor countries. This is a more complex and interesting argument, but there is little evidence or support for any of its variants. For example, the 13 There are, however, sound empirical reasons to be concerned about causality in the relationship between the security of property rights and economic growth. Among these, measurement error ­ observer bias that boosts assessments of the security of property rights in fast-growing countries ­ is the most important. 18 citizens of poor countries, being poor, might discount the future more heavily and thus be more willing to incur the future costs of expropriation in return for the short-run benefits. While persuasive in principle, this argument is inconsistent with the observed willingness of the poor to take actions with distant payoffs and high current costs, such as investing heavily in the education of their children. Poor countries might also exhibit higher income inequality, and therefore more policy volatility or willingness to expropriate (as, for example, Knack and Keefer 1997a argue). However, it is well-known that income inequality and income exhibit an inverted U-shaped relationship: inequality is low at both low and high levels of income per capita, and higher over middle ranges. Such a relationship cannot generate a spurious correlation between growth and the security of property rights unless slow-growing countries are on average richer and fast-growing countries on average poorer. Evidence from the growth literature suggests that the reverse may be true, however. A related literature argues that as countries grow wealthier, more effort and resources are spent to define and protect property rights. The insights of this literature are important in their own right, but they cannot be taken to support the argument that growth leads to secure property rights. The arguments here are simply that as countries grow, assets that previously held little value and were not worth protecting become more valuable, increasing the payoffs to defining ownership rights to them carefully.14 The governance literature focuses on the insecurity of property rights to valuable assets. 14 Rangeland in the American West had this characteristic ­ not so valuable prior to the arrival of the railroads, but very valuable and worth fencing after their arrival. Esther Boserup (1965) first argued that population pressure increases the value of land, leading to greater efforts to protect rights to land. Anderson and Hill (1990) show that the conditions under which investors (farmers, in this case) are allowed to establish property rights also has a significant effect on productivity. However, none of these studies show (nor are they intended to show) that growth makes insecure property rights secure. They demonstrate instead that growth makes it 19 Even if reverse causality is not an issue in the relationship between property rights security and economic growth, it is possible that other factors, omitted from most analyses, might simultaneously drive down the security of property rights and economic growth, yielding a spurious correlation between the latter two. Social polarization, political instability, the absence of political checks and balances, or unfortunate geographic location all might explain both insecure property rights and slow economic growth. However, for these omitted variables to create a problem of causality, it must also be the case that their effect on economic growth is largely through channels other than the security of property rights. This seems unlikely. If political instability or social polarization matter for growth, it is most reasonable to assume that they matter exactly because they drive down the security of property rights. These are in fact the arguments in Acemoglu and Robinson (2001), Keefer and Knack (2002a) and Svensson (1998). Such omitted factors do not cause a spurious relationship between the security of property rights and growth. On the contrary, they precipitate the causal chain that runs from the first to the second. Predictability and credibility Some contributions to the governance literature refer not only to the security of property rights but to the closely related but somewhat broader concept of "government credibility." This is the case with Knack and Keefer (1995), for example. Only credible governments can assure investors that their assets are safe from expropriation. In addition, however, governments that are not credible cannot elicit any actions from citizens or investors in return for any future promise of government action. In times of worthwhile to define previously worthless property rights. Worthless property rights are not, however, at the heart of the debate linking the security of property rights to economic development. 20 fiscal crisis, for example, the non-credible government cannot offer future rewards to constituencies in exchange for current cutbacks. These governments cannot use any policy instrument to stimulate investment short of outright subsidies to investors that compensate them for their risk. As in the case of property rights, an important definitional confusion arises in the area of credibility, particularly in empirical applications. Attempts to capture credibility issues in surveys tend to rely (as in many World Bank surveys, for example) on questions to respondents on their perceptions of the "predictability of decision making". At first glance, this seems a perfectly reasonable approach to governance: no one likes unpredictable decision making, and unpredictable and arbitrary decision making are surely pillars of bad governance. The problem is not that unpredictable decision making is good, however, but rather that it is not necessarily so bad and does not capture the sort of asymmetrical and unbounded threat that is implied by the lack of government credibility. Strictly speaking, unpredictability is a "mean-preserving" increase in the variance of future government policies. That is, unpredictability leaves the average or expected policy unchanged, but increases the range of potential policies that could emerge in the future. Predictability should have only a modest effect on economic outcomes (see Aizenman 1998, for example). Assuming plausible parameters and an often-used functional form for investor utility, it is easy to show that a moderately risk-averse investor who would demand, say, a 10 percent rate of return for a sure investment would demand just 11 to 12 percent for an investment with a payoff that with 50 percent probability returns 50 percent less than the sure payoff, and with 50 percent probability 21 returns 50 percent more. This sort of risk premium cannot explain the differential investment flows into Russia and Poland, or Zaire and Botswana. Credibility, on the other hand, is not mean-preserving: Instead, it is an asymmetric risk that emerges whenever investors or households make illiquid investments, from which governments can subsequently benefit by devaluing.15 Bureaucratic efficiency and honesty There is a long line of research, stemming from Max Weber and extending to many current discussions of state capacity or the strength of states, such as Evans (1995), arguing that the actions of bureaucracies have an independent effect on economic development. However, although it is well-known that bureaucratic quality is lower and corruption higher in poor countries, the inference that these phenomena cause poverty and slow growth is more difficult to draw. Of course, to the extent that government policies are ideal and government spending ideally allocated, then any bureaucratic breakdown must lower social welfare and, to the extent that the policies and spending are connected to economic growth, slow growth. What, however, if the policies and spending decisions are entirely deficient? The classic conundrum is therefore the following: are societies better off when honest bureaucracies rigorously and predictably apply abusive regulations of firms and individuals or, given the existence of such laws and regulations, would citizens prefer that they be implemented by corrupt bureaucracies? In the context of corruption, Leff (1964) was the first of many to make the argument that corruption could be beneficial if it 15 If there is a high correlation between the answer to the simple question ("how predictable is government decision making?") and the nuanced issues of credibility ("How easy would it be for a government official to make a decision or take some action that would reduce your profits by 50%?"), then this strong theoretical distinction is less important for empirical work. 22 allowed citizens to circumvent welfare-reducing laws and regulations. More recent analyses have emphasized that arguments such as Leff's assume that bureaucracies have no influence on the content of the regulations that they implement. If they do, though, the potential for lucrative bribes would spur bureaucracies to generate abusive regulations that maximize bribes. Kaufmann and Wei (1999), have argued and provided evidence that in this more dynamic world, bureaucrats react to the opportunities for corruption by actually increasing red tape and reducing their (bribe-free) performance. This later literature, in turn, relies on the assumption that bureaucracies enjoy significant discretionary authority. It is a characteristic of the governance literature that, although it gives apparently great weight to political processes in its emphasis on democracy, voice and accountability, it tends to analyze bureaucratic performance in isolation and to ignore the intimate relationship of bureaucrats and politicians. If one were sure that bureaucrats could exploit information and other advantages to act independently and even in opposition to the objectives of their political superiors, this assumption would be entirely appropriate. One could immediately begin to examine the effects of corruption or bureaucratic incompetence on outcomes. However, research into the delegation of authority by political actors to bureaucracies has made clear that politicians expend considerable effort to rein in deviant bureaucrats and to circumvent or compel recalcitrant ones. Weingast and Moran (1983) and Keefer and Stasavage (2003) show systematic differences in the behavior of bureaucracies (the Federal Trade Commission and central banks, respectively) depending on the partisan control of government and the number of political checks and balances (respectively). Epstein and O'Halloran (1999) document at 23 book-length the limits of bureaucratic information advantages over politicians and the many ways that politicians limit them. McCubbins, Noll and Weingast (1987) have described in detail the constraints on bureaucratic discretion imposed by the US Congress through the Administrative Procedures Act and other vehicles. Kang (2002), while not relying on this literature himself, still demonstrates the intimate link between political imperatives and the nature of corruption in South Korea and the Philippines. In view of this literature, it seems likely that the observed effects of corruption or bureaucratic inefficiency on economic outcomes are more properly attributed to politician behavior that undermines both outcomes and bureaucratic performance, rather than to the specifics of bureaucratic organization. This logic raises doubts about the causal relationship between growth or development and the characteristics of a country's bureaucracy. First, political incentives and politician actions are likely to generate both bad policies and a neglected, inefficient and/or corrupt bureaucracy. Second, it is likely that the bad policies, as much as or more than the characteristics of the bureaucracy, drive bad development outcomes, potentially giving rise to a spurious association between bureaucratic characteristics and development. This is not to say that the bureaucracy is unimportant. Even if politicians substantially constrain bureaucratic discretion, residual discretion could nevertheless have meaningful economic consequences. However, there is substantial justification ­ not least of which are the numerous examples of well-oiled bureaucracies implementing catastrophic policies ­ to demand that any examination of this issue must be careful to 24 control for the possibility that observed relationships between bureaucratic performance and economic growth or development might be spurious. Voice and accountability as governance A vast literature considers the importance of politics and political institutions in the process of economic development, covering topics ranging from regime type to political instability to clientelism to the institutions of electoral competition and political decision making. The subset of that literature that falls under the governance rubric centers on voice, accountability and democracy. This focus dates at least to Albert Hirschman (1970), who introduced "voice" and "exit" into the development lexicon. The emphasis on voice and accountability rests on a natural and persuasive logic: governments that pay little attention to citizen concerns ­ because citizens are unable to voice them or are unable to sanction governments that ignore them ­ are less likely to pursue policies that further social welfare. Less clear are the specific attributes of political and social organization in a country that improve voice and accountability. The governance literature has explored systematically only one aspect of political organization that bears on voice and accountability, regime type, or democratic versus autocratic government. The presumption is that democracy is the only regime type that affords voice and accountability to citizens on a regular basis. Democracy should therefore promote economic development. In fact, most of the research into democracy and economic growth is empirical and yields ambiguous results. Przeworski and Limongi (1993) underline the theoretical obstacles that might be responsible for this ambiguity. They begin by pointing out that much of the democracy literature argues that democracies raise growth rates by providing stronger guarantees for the security of property rights. But they argue that nothing about 25 democratic decision making, seen as majority rule, makes expropriation more difficult, while long-lived dictatorships might provide even stronger protection for property rights.16 Their critique of the democracy and growth literature follows both the governance work on voice and accountability and the literature on democracy in treating the state as a unitary actor. Such analyses abstract from the substantial variation among polities in their political and electoral institutions, the dynamics of electoral competition, and differences in the constellations of interest groups across countries.17 Integrating results and tools from the vast political economy literature that focuses on precisely these details is likely to provide fertile ground for assessing the underlying institutions that influence voice, accountability and, ultimately, economic development. The final section of this chapter discusses the significant payoffs to future research that bridges the governance literature with new developments in political economy. Evidence of the effects of governance on economic development Evidence from cross-country statistical research amply supports the importance of secure property rights, the rule of law, or the credibility of government for growth (Knack and Keefer 1995, Acemoglu, et al. 2001, Rodrik, et al. 2002) or per capita incomes (Hall and Jones 1999). However, as Weyland argues, there are large potential econometric problems that emerge in this literature. For example, the empirical governance measures are usually subjective, introducing noise and, possibly, bias into the resulting estimates. 16 Przeworski, et al. (2000) introduce and provide substantial evidence for several novel variations on the theme of democracy and growth, showing in particular (though not explaining) that population growth is faster in dictatorships; controlling for this, per capita income growth is approximately the same in both dictatorships and democracies. 17 They summarize democratization as giving the median voter control over policy, leading to a shift in distribution towards the median voter and a consequent disruption, rather than reinforcement, of property rights. 26 Nevertheless, they represent a quantum leap in the ability to bring statistical analysis to bear on topics that were previously investigated only in theory or case studies, a leap that can be measured in terms of the sharp increase in research on governance-related topics that followed the introduction of these measures into the literature. Some of this literature is reviewed here. Empirical tests of the effects of government credibility/secure property rights on development With respect to the issue of property rights or the rule of law, fairly exhaustive attempts have been made to control for empirical difficulties ranging from endogeneity to measurement error. Results documenting the importance of secure property rights for growth have been robust to causality testing (Calderón and Chong 2000), ingenious instruments and historical investigation (Acemoglu, Johnson and Robinson 2001), and to the substitution of income for growth as the dependent variable (Hall and Jones 1999). Although there is still room for skepticism, it is nevertheless true that the influence of secure property rights has withstood an unusually large amount of scrutiny. The security of property rights has also been linked to development phenomena other than economic growth. Dollar and Kraay (2002) consider the impact of secure property rights on inequality, via growth. They find that the security of property rights raises the incomes of the rich and poor equally. Using beginning of period measures of property rights and later measures of inequality, Knack (2002) finds that secure property rights are actually progressive, raising the incomes of the poor faster than those of the rich. In a much different application, Deacon (1999) has found that the security of property rights is negatively related to rates of deforestation. 27 The hypothesis that secure property rights accelerate economic development is supported by quantitative approaches other than those employing cross-country data, as well. Natural experiments involving property rights include work by Feder (1993) on the consequences of land titling for farmers in Thailand; Jiménez (1984) on the effects of titling on investment in and prices of homes in Manila; Alston, Libecap and Mueller (1999) on the effects of property rights security on farm productivity and deforestation in the Brazilian Amazon; and O'Rourke (1999) on the effects of land tenure insecurity on the adoption of efficient butter production methods in Ireland contrasted with Denmark. One caveat that is useful to recall in reviewing the empirical work on property rights and growth is that this work often contains broader statements about the importance of "institutions" more generally. Knack and Keefer (1995), Acemoglu, Johnson and Robinson (2001), and Rodrik, Subramanian and Trebbi (2002) all argue for the dominant effect of "institutions" on growth, but all use measures of the security of property rights as their measures of institutions. In the case of Knack and Keefer (1995), despite the use of the word "institutions" in the title of the paper, both the hypotheses and the tests relate to the security of property rights and the rule of law. In other cases, however, the property rights variables are meant to stand for broader, less well-specified institutional concepts. The thrust of these papers is that "institutions" matter, not that "secure property rights" matter. This would be entirely benign if all of the institutional and governance forces that promote development were captured by measures of the security of property rights. This does not seem to be the case, however ­ nor would one expect it to be in view of the heterogeneity in the theoretical links between the different components of governance and economic outcomes. 28 Empirical research into corruption and bureaucratic quality The conclusions from the empirical work on corruption and bureaucratic quality are broadly similar to the work on property rights. In the first and prototypical empirical contribution, Mauro (1995) shows that corruption reduces growth, though Li, Xu and Zou (2000) show that taking inequality into account reduces the estimated effect of corruption. Gupta, Davoodi and Alonso-Terme (2002) present evidence that corruption exacerbates income inequality and poverty. Mo (2001) documents a causal chain linking higher corruption to lower growth through reduced investment in human and private capital. Evans and Rauch (2000) present quantitative evidence supportive of the "embedded autonomy" hypothesis of Evans and controlling for income, education and ethnic and linguistic diversity, find that some characteristics of an "autonomous" bureaucracy (especially meritocratic recruitment) predicts the assessments of property rights security and bureaucratic quality used elsewhere in the literature. It is, however, more difficult to draw firm conclusions about the effect of corruption and bureaucratic quality on economic development. On the one hand, in the cross-country work exemplified by Mauro (1995), the corruption variables exhibit considerable overlap with the measures used in the property rights literature. Mauro's variable is highly correlated, for example, with the property rights variable of Knack and Keefer, which is consistent with endogeneity problems to which the analysis of corruption is likely exposed.18 On the other hand, as the earlier discussion foreshadows, the more fundamental difficulty in drawing inferences about the impact of corruption and bureaucratic quality 18Still, it is fair to point out, that among the variables that make up these risk guide indices, the lowest inter- variable correlation is with corruption, suggesting that the risk guide indicators are able to differentiate at least somewhat between corruption and other governance problems. 29 on development is their greater vulnerability to causality problems. Corruption in government is sensitive to the incentives of the political leadership, which are typically not controlled for in corruption work and are sure to have a significant independent effect on government performance and economic development. These causality problems emerge in at least two specific ways. First, governments often put into place distortionary policies that directly slow economic development but also give rise to corruption. Broadman and Recanatini (2002) use evidence from surveys of businesspeople in transition countries regarding corruption and the effects of regulation to show that high regulatory barriers to entry and soft budget constraints on firms are both conducive to corruption. More optimistically, Cheng, Haggard and Kang (1998) argue that well-functioning bureaucracies in South Korea and Taiwan were the result of conscious decisions by political actors who, in turn, were motivated by a desire to avert past crises or to deter aggression by large neighbors. Second, corruption can emerge from other governance problems that also directly limit development. Keefer and Knack (2002b) present a model and evidence that if political decision makers find it difficult or costly to guarantee the security of property rights of private investors, they will change their policies accordingly. In particular, they will suppress productive public investment and increase rent-seeking or corruption. The evidence in this paper suggests that countries with less credible governments (countries with less secure property rights) spend 60 percent more on public investment as a fraction of national income than countries with more secure property rights and that the difference is due the effect of insecure property rights on incentives to divert rents.19 No 19 Political Risk Services generates the International Country Risk Guide, assessments of country risk widely used by foreign investors. Two variables in the ICRG, bureaucratic quality and corruption, comprise the 30 investigations of the effect of corruption and bureaucratic quality on economic development take such endogeneity issues into account either at the level of theory or empirical testing. Voice, accountability and democracy: the evidence There are no agreed indicators of the concepts "voice" and "accountability", either in theory or in the empirical literature. The literature employs proxies ranging from Freedom House indicators of political freedoms and civil liberties (found by Scully 1988 to be predictors of growth), to structural indicators of democracy ­ whether there are competitive elections and the executive is constrained by a legislature. With respect to the Freedom House indicators, the empirical record is reasonably convincing that these variables are important for development. Like governance itself, these variables are themselves multi-dimensional. This makes it difficult to determine which concrete and objectively observable characteristics can be traced specifically to notions of voice and accountability, and which are specifically responsible for observed changes in development outcomes. The democracy literature, relying on more structural and objective characteristics of polities, yields much more ambiguous results. One source of confusion is that democratic countries are both richer and better performing. This naturally gives rise to the difficult question: are countries better performing because they are rich, or because they are democratic? For example, as we would expect, corruption and rent-seeking are higher in countries lacking fully competitive elections, as are the risks of expropriation.20 However, countries with competitive elections are also substantially richer than other measure of "bureaucratic performance" used in Keefer and Knack (2002). 20This comparison is between countries that have competitive legislative and executive elections (LIEC and EIEC equal to seven, from the Database on Political Institutions, Beck, et al., 2001) and countries that do not. 31 countries, even excluding the richest 20 democracies from the comparison (in 2000, excluding the richest 20, 97 countries with competitive elections had purchasing power parity adjusted per capita incomes of $5,470, compared to $3,121 for the countries that did not have fully competitive elections). In part because of this, a large literature that always controls for per capita incomes at the beginning of the period finds only an ambiguous relationship between democracy and growth. Many of the income gains that democracy generates may already be embedded in the higher initial per capita incomes of democratic countries.21 A number of investigations use more global indicators to test broader notions of governance or institutions on economic development. Easterly and Levine (2002), also engaging the debate on the role of institutions versus geography in economic development, use the global governance index of Kaufmann, et al. rather than more narrow property rights and rule of law indicators. The broad governance index has the advantage of including more dimensions of a country's institutional performance. However, the aggregate measure exacerbates the problem of interpretation. Provided the coefficient estimates are significant, one can only say that some factors related to governance are positively associated with growth. One cannot say, however, which factors those are, nor even be sure that some other factors do not also have a negative impact on growth. 21Of course, Przeworski and Limongi (1993) would argue that there is in fact no theoretical case for arguing in favor of the superior economic or growth performance of democratic countries. Among the research documenting the empirical ambiguity are De Haan and Siermann 1995 and, especially, Przeworski, et al. (2000). On the other hand, Quinn and Woolley (2001) argue that the effect of democracy is to make the economy less volatile, reflecting the preferences of citizens for lower risk (which, in democracies, they are better able to express). 32 The research frontier: identifying the determinants of good governance While governance research has made convincing contributions to the study of development, crucial questions remain. Some of these have to do with endogeneity issues raised by much of the governance literature. For example, does corruption suppress development or is corruption just symptomatic of a deeper problem in the political environment of a country? Understanding the determinants of good governance can help answer this question. However, regardless of the endogeneity and causality issues that are preoccupations of the governance literature, the determinants of good governance ­ of secure property rights, voice and accountability, or honest and efficient bureaucratic behavior ­ are key issues in their own right. Intellectually, they are some of the most challenging in the social sciences. From a policy perspective, they are essential to understand if one is to move forward in developing sustainable and effective reforms for countries suffering from, for example, insecure property rights. There are two important political economy research efforts that relate to the sources of good governance. One relates the structure of political institutions and competition, including the information of voters, the credibility of political competitors and the underpinnings of intra-party competition, to broad decisions of government, particularly related to fiscal allocations but extending to corruption. The second focuses more specifically on the conditions under which countries exhibit secure property rights. The political economy of government spending and corruption Recent important advances have brought great rigor to our understanding of the impact of political and electoral institutions, credibility and voter information on government incentives regarding fiscal policy, including rent-seeking (the textbook by Persson and Tabellini 2000 provides a wide-ranging and complete summary). This 33 literature focuses on the information of voters, the ability of governments to make credible pre-electoral promises to voters, and the details of the institutions of government decision making (parliamentary or presidential; if presidential, is it the president or the legislature that prepares the budget? What are the amendment powers of the respective entities?). The power of the analysis lies not so much in the accuracy of its assumptions about the political process, but rather in the coherent way in which the analysis marshals the many influences, ranging from institutions to information, that act on political decision makers' policy making incentives. One particular emphasis in Persson and Tabellini (2000) is the role of political and electoral institutions on political incentives. For example, they predict that financing of public goods (those that benefit the whole population) relative to more targeted goods is higher in parliamentary democracies. In fact, among 24 parliamentary democracies for which data are available, the ratio of education (broad and relatively untargeted) to public investment spending (narrow and relatively targeted to specific voters) is about 20 percentage points higher than in 18 presidential democracies.22 This has direct implications for governance work, which is concerned precisely with using government to efficiently improve the welfare of the average citizen ­ a goal that is more difficult to achieve when political institutions give policy makers incentives to provide inefficient targeted goods that benefit narrow slices of the population. Models from this literature also predict that corruption (rent-seeking) should be higher under presidential systems and under non-proportional electoral systems with larger district magnitudes (legislative seats per electoral district). Empirical research has 22 Using political measures for the Database on Political Institutions, Beck et al. (2001). 34 provided support for these predictions, as well. Persson, Tabellini and Trebbi (2002) find that electoral systems with non-proportional electoral rules and larger district magnitudes discourage corruption. Lederman, Loayza and Soares (2001) conclude that presidential systems are more corrupt. The predictions regarding political and electoral institutions and their effects on policy outcomes or governance are sensitive to the underlying assumptions regarding how well-informed voters are, how ideologically divided they are, and how credible political promises to voters are prior to elections. Direct examination of these assumptions is a natural and important direction for future research. Already, however, research has provided early insights into the independent role of many of these assumptions on development outcomes. Besley and Burgess (2001), for example, report an illuminating natural experiment about the role of information in political decision making. They show that in Indian states with greater literacy, electoral turnout and newspaper circulation (controlling for per capita incomes), state governments were more likely to respond to food shortages via the public distribution of food. Their work takes just one of many potential information problems as a point of departure. They leave aside the information issues prominent in Persson and Tabellini (2002), particularly the effects of voter and politician uncertainty about their respective ideological biases. They also apply their analysis to one particular policy problem, necessarily abstracting from, for example, the security of property rights. Much of this work follows the "rational choice" or "new political economy" style of analysis ­ bringing more quantitative approaches to the study of individual actors and 35 their incentives and largely ignoring historical or cultural differences that might influence outcomes. The comparative politics literature is full of valuable contributions that follow the reverse strategy, differentiating regime types by referring to historical and broad characteristics rather than to the incentives of the actors that run them. Remmer (1986) categorizes democracies as "exclusionary" and inclusive, for example, and Bratton and van de Walle (1994) argue that African democracies are different because they have their roots in neopatrimonial forms of government, where the executive maintains authority by dispensing personal patronage. In neither case are the authors explicit about the underlying characteristics of political competition or political institutions that permit elected officials to be exclusionary or neo-patrimonial. Whereas Bratton and van de Walle take as a characteristic feature of regimes the personal dispensation of targeted or private goods to individuals or small groups, the literature summarized in Persson and Tabellini (2000) seeks to explain the reliance on targeted public spending. Efforts to bridge these two literatures can yield critical insights into the nature of economic development. For example, the new political economy literature regularly makes explicit and often different assumptions about the credibility of political promises, but it rarely analyzes variation in credibility across countries as a key determinant of differences in development outcomes. Keefer (2002) makes an attempt in this direction, arguing that clientelism prevails when competing politicians can make credible pre- election promises to only a few people and dissipates when their promises are credible to the whole country. Lacking institutionalized bases for credibility (such as long-standing political parties with well-identified ideological platforms), political competitors can therefore only make credible promises to a small fraction of the population, leading to a 36 preference for targeted, personalized and inefficient spending rather than efficient and broad-based public good provision.23 This argument is rooted in the detailed observations of clientelism (e.g. by Scott 1972) and of the patterns of political recruitment in relatively clientelist democracies (e.g., the work of Krishna (no date) on candidate recruitment in India), but uses the tools summarized in Persson and Tabellini (2000) to explain why these patterns exist in some countries and not others and to sort out their policy implications. The value of the effort can be measured by its success in explaining important puzzling phenomena. There are two, in particular, to which the conclusions in Keefer (2000) seem to apply. First, to the extent that political actors in poor countries are less able to make credible promises to voters, they are more likely to spend on targeted and narrow constituencies, such as most public investments, than untargeted constituencies, as is the case with most education spending. In fact, in 1997, the ratio of education to public investment was 70 percent higher in the 12 richest parliamentary democracies than in the 12 poorest (that is, holding institutions constant). At the same time, the 12 poorest parliamentary democracies allocated resources in almost the same way as 14 other countries that could not be considered functioning democracies at all (and for which data are available on education spending). Second, as Bratton and van de Walle have argued, history seems to matter in the development of democracies. Bratton and van de Walle argue that there is a relationship 23 Though the credibility approach to clientelism is arguably most consistent with the literature on clientelism and neo-patrimonialism, there are two important alternative theories. Robinson and Verdier (2002) argue that clientelism emerges when politicians can make no credible promises to anyone, but have a high regard for the welfare of some small group of citizens (e.g., members of the same tribe or citizens from the same region). Medina and Stokes (2002) assume politicians are always credible, but that clientelism emerges when some politicians have monopoly control over assets that voters value. 37 between the nature of the pre-democratic regime and the subsequent possibilities for consolidating democracy, although they are less clear about the underlying causal relationship. Keefer (2002) argues that the pre-democratic period determines whether political competitors in a democracy have credible policy stances on a wide range of issues. Evidence for this comes from the contrasting experience of British democracy after the massive expansion of the franchise in the early 19th century and the evolution of democracy in the Dominican Republic following the death of Rafael Trujillo. The first succeeded because of a history of political competition between relatively coherent and therefore credible political alternatives; the second has proceeded haltingly because of the absence or ruthless suppression of political competition prior to democracy. The personalized relationships that characterize clientelism are found in other contexts as well. Heilbrunn's (2002) examination of much-publicized corruption episodes in France over the 1970s - 90s demonstrates the role of personal relationships ­ networks ­ among key decision makers. These informal networks circumvented or short- circuited apparently robust institutional checks and balances meant to control corrupt behavior. As a result, although one party was consistently associated with massive campaign spending scandals, including large procurement bribes associated with the purchase of warships by the Taiwanese and the diversion of money from the state oil company to party coffers and to the bank account of the chief judge of the Supreme Court, the other party refrained from pushing for prosecution. Personal, cross-party ties ­ much as in many developing countries ­ seem to explain this. Heilbrunn (2002) argues that the circumstances that allowed these personalized ties to overwhelm the institutional checks and balances: as long as members 38 of the network controlled sufficient levers of power, network members could be sure that the rewards of loyalty to other network members would exceed the potential costs of subverting the institutional checks and balances. The unraveling of these cases provides further evidence for the importance of networks: prosecution of these officials was eventually undertaken, but only at the initiative of a Norwegian-borne female prosecutor, someone outside the networks documented by Heilbrunn. In sum, the literature on the political roots of good governance offers promising directions for future research. The most promising seem to relate to the specific conditions of political decision making and electoral decision making, specifically to information, credibility and the nature of inter-personal relationships among politicians. The political economy of secure property rights The literature on the determinants of secure property rights emphasizes especially the role of political checks and balances, but also social characteristics such as polarization and social capital. Substantial puzzles remain, however. For example, property rights are insecure in many countries with formal institutional checks and balances, secure in others lacking them. North and Weingast (1989) ask why the interest rates charged by Dutch lenders to the English Crown fell after the Glorious Revolution and argue that the introduction of checks and balances in government (a more powerful parliament) reduced the risk to lenders that the English Crown would renege on its contractual obligations. Stasavage (2003) and others have pointed out a substantial time lag (as much as sixty years) between the Revolution and the decline in interest rates, however. He argues that the introduction of additional institutional checks is insufficient to secure property rights. He shows that the time lag can be explained by the slow evolution of the preferences of the 39 different actors who controlled the newly instituted checks and balances. To the extent that the preferences of at least one "veto player" were aligned with those of lenders to the English Crown, checks and balances secured lender contractual rights; this did not occur, however, until some years after the institutional change took place.24 Keefer and Stasavage (2002) also find that both preferences and institutional veto points matter in ensuring credibility with respect to monetary policy. The literature is convincing that checks and balances are important to government credibility and secure property rights, but suggests as well that they are neither necessary nor sufficient. Essays in Boortz and Haber (2002), about the ability of the Mexican dictator Porfirio Diaz to entice considerable investment in the absence of institutional checks and balances, make it clear that checks and balances may not be necessary for government credibility or secure property rights. In cross-country comparisons, checks and balances have some effect on the security of property rights, but the effect is not robust to the presence of controls. There are straightforward econometric difficulties that can explain the weak statistical relationship: the inevitable correlation of other explanatory variables, such as per capita income, with checks and balances and the fact that empirical variables used to represent checks and balances represent more accurately the number of institutional veto gates, but less accurately the preference alignments of the veto players who occupy those veto gates. However, beyond these difficulties, it is clear as well that the conditions under which non-democracies can establish credible 24 Hallerberg (this volume) reviews similar arguments regarding policy stability. Keefer and Stasavage (2002) show, however, that policy stability and policy credibility are not the same. Policy can be stable (e.g., policy will always equal the median voter's preferred outcome in a one dimensional, simple model of policy making) but still not be credible (since, conditional on the response of citizens to the median voter's decision in the first period, the median voter may still gain from reneging or changing the policy, in the next period). 40 commitments and some democracies cannot remains an open and important area of inquiry. The work of Clague, et al. (1996) deepens the property rights puzzle, highlighting the role of regime longevity. They find that long-lived autocrats offer as much or more protection to property rights than short-lived democracies, though long-lived democracies are the most secure of all. This conclusion reinforces the idea that the protection of property rights depends on more than formal institutional rules, but it also guides future research towards a more specific puzzle: what changes over time in democracies, such that older democracies provide more secure property rights than younger ones? Keefer (2002) suggests that young democracies are more vulnerable to clientelism and, by implication, restricting to clients the fraction of investors to whom the government can make credible promises about the security of property rights. The literature has yet to rigorously link clientelism and the security of property rights, however. Social explanations for the security of property rights also abound. Knack and Keefer (1997b) find that measures of social capital ­ essentially, measures of the extent to which people in different countries believe that others in their country are trustworthy ­ are positively associated with the security of property and contract rights. Naturally, causality problems emerge here: do people trust each other because there institutions make property rights and contractual commitments reliable, or the other way around? In either direction, however, there are important implications for development that have yet to be explored. Polarization in society might also undermine property rights security ­ the argument being either that the more polarized are citizens, the less likely they are to 41 respect the rights of others, or more in the vein of social choice, the more polarized are citizens, the more volatile is policy likely to be. Svensson (1998) has found inequality and property rights to be inversely associated, and offers econometric evidence that the effect is from the first to the second. Knack and Keefer (2002a) show that other forms of polarization ­ ethnic and linguistic ­ suppress the security of property rights, and it is in part through this channel that they affect growth. Acemoglu and Robinson (2001) argue that wealth inequality can increase political instability and, thereby, the insecurity of property rights. Keefer and Knack (2001) find evidence that income inequality reduces insecurity, but no support for the role of political instability. All of these contributions still leave up in the air the specific channels through which polarization undermines the security of property rights, however. Though following similar empirical strategies, for example, Svensson (1998) and Knack and Keefer (2002a) propose completely different channels: Svensson argues that property rights security are the result of investments by government ­ a fiscal decision ­ while Knack and Keefer model the security of property rights as the propensity of government decisions to change dramatically and unpredictably. In sum, then, large questions about the determinants of property rights remain. There is evidence that long-running democracies promote property rights, broadly defined, but we do not know which characteristics of long-running democracies endow them with this characteristic. Nor do we have micro-level evidence that more systematically outlines the actions that are open to decision makers in short-lived democracies that are foreclosed to those in long-lived democracies. One might interpret the observation about clientelism as implying that pre-democratic forms of political 42 exchange survive the introduction of democratic institutions ­ but we do not know how or under what conditions they disappear. A promising line of research is surely to link these observations with the vast literature on the dynamics of political party evolution and the conditions under which political parties can make credible promises to voters. Conclusion: How can we improve governance? Governance reform prospects and strategies depend significantly on whether one believes that flaws in the state apparatus (e.g., public administration) lie at the heart of governance failures, or whether one believes that they are more deeply embedded in the political and social dynamics of a country. If the former, then significant opportunities for reform open up. For example, corruption can be addressed by reforms to government financial management, procurement and audit systems. Voice and accountability can be addressed by requiring bureaucracies to open up the rule making process (the regulatory process) to more popular participation. Bureaucratic quality can be improved by boosting pay and strengthening meritocratic recruitment procedures. If, on the other hand, governance failures are more deeply rooted in the incentives of political actors, these reforms may not translate into significant change in the way government operates ­ in the security of property rights, in the quality of regulation, or in the services provided by governments to citizens. This does not imply, however, that reform is impossible or that there is no role for outside assistance. Instead, it implies that reform must be structured to address or adapt to the underlying difficulties in the relationship between voters and politicians. Those outside the political process can potentially change the political equilibrium by providing relevant information about candidate performance to voters, verifying or debunking political claims of responsibility 43 for good outcomes, and the mobilization of voters around service delivery issues (that is, making it more credible to politicians that voters whose apathy or support they took for granted might actually vote against them on the basis of service delivery or other dimensions of government performance that were not previously relevant). There is unlikely to be a generic blueprint good for all reforms in all countries. However, the literature is moving toward the identification of key, concrete political obstacles to good governance that, like the extent of voter information, are amenable to change. References Acemoglu, Daron and James Robinson (2001). "A Theory of Political Transitions." American Economic Review 91:4, 938-963 (September). Acemoglu, Daron, Simon Johnson and James Robinson (2001). "The Colonial Origins of Comparative Development: An Empirical Investigation." American Economic Review 91, 1369-1401. Aizenman, Joshua (1995). "Investment in New Activities and the Welfare Cost of Uncertainty." Journal of Development Economics 52:2, 259-77 (April). Alston, Lee, Gary Libecap and Bernardo Mueller (1999). Titles, conflict and land use: The development of property rights and land reform on the Brazilian Amazon frontier. Ann Arbor: University of Michigan Press. Anderson, Terry L. and Peter J. Hill (1990). "Race for property rights." Journal of Law and Economics 33: 177-97 (April). Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer and Patrick Walsh (2001). "New Tools in Comparative Political Economy: The Database of Political Institutions." World Bank Economic Review, 15: 1, 165-176. Besley, Timothy and Robin Burgess (2001). "Political Agency, Government Responsiveness and the Role of the Media." European Economic Review 45:4-6, 629-40 (May). Boortz, Jeffrey L. and Stephen Haber (2002). The Mexican Economy, 1870-1930: Essays on the economic history of institutions, revolution and growth. Stanford, Ca.: Stanford University Press. Boserup, Ester (1965). The conditions of agricultural growth: the economics of agrarian change under population pressure. New York: Aldine Publishing Co. Bratton, Michael and Nicholas van de Walle (1994). "Neopatrimonial Regimes and Political Transitions in Africa." World Politics 46:4, 453-489 (July). Broadman, Harry and Recanatini, Francesca (2002). "Corruption and Policy: Back to the Roots." Journal of Policy Reform 5:1, 37-49 (March). Buchanan, James M., Robert D. Tollison, and Gordon Tullock, eds. (1980). Toward a Theory of the Rent-Seeking Society. College Station, TX: Texas A & M University Press. Calderón, César and Alberto Chong (2000). "Causality and Feedback between Institutional Measures and Economic Growth." Economics and Politics 12:1, 69 ­ 81 (March). Cheng, Tun-Jen, Stephan Haggard and David Kang (1998). "Institutions and growth in Korea and Taiwan: The Bureaucracy." Journal of Development Studies 34:6, 87-111 (August). 45 Clague, Christopher, Philip Keefer, Stephen Knack and Mancur Olson (1996). "Property and Contract Rights under Democracy and Dictatorship." The Journal of Economic Growth 1:2, 243-276 (June). Deacon, Robert T. "Deforestation and Ownership: Evidence from Historical Accounts and Contemporary Data." Land Economics 75:3, 341-59 (August). Dollar, David and Aart Kraay (2002). "Growth is Good for the Poor." Journal of Economic Growth 7(3): 193-225. Easterly, William and Ross Levine (2002). "Tropics, Germs and Crops: How Endowments Affect Economic Development." National Bureau of Economic Research Working Paper 9106 (August). Epstein, David and Sharyn O'Halloran 1999. Delegating powers : a transaction cost politics approach to policy making under separate powers. New York : Cambridge University Press. Evans, Peter (1989). "Predatory, Developmental, and Other Apparatuses: A Comparative Political Economy Perspective." Sociological Forum 4:4, 561-87. _____ (1995). Embedded Autonomy: States and Industrial Transformation. Princeton: Princeton University Press. Feder, Gershon (1993). "The Economics of Land and Titling in Thailand." In Karla Hoff, Avishay Braverman and Joseph E. Stiglitz, eds., The economics of rural organization: Theory, practice, and policy, pages 259-68. New York: Oxford University Press. de Haan, Jakob and C.L.J. Siermann (1995). "New Evidence on the Relationship between Democracy and Economic Growth." Public Choice 86: 175-98. Gupta, Sanjeev, Hamid Davoodi and Rosa Alonso-Terme (2002). "Does Corruption Affect Income Inequality and Poverty." Economics of Governance 3:1, 23-45. Hall, Robert E. and Charles I. Jones (1999). "Why Do Some Countries Produce So Much More Output per Worker than Others?" Quarterly Journal of Economics 114:1, 83-116 (February). Heilbrunn, John (2002). "Oil and Water? Elite Politicians and Corruption in France." Mimeo, Colorado School of Mines. Hirschman, Albert (1970). Exit, Voice and Loyalty: Responses to Decline in Firms, Organizations and States. Cambridge, Massachusetts: Harvard University Press. Jímenez, Emmanuel (1984). "Tenure security and urban squatting." Review of Economics and Statistics 66:4, 556-67 (November). Kang, David (2002). Crony Capitalism: Corruption and Development in South Korea and the Philippines. New York: Cambridge University Press. 46 Kaufmann, Daniel and Aart Kraay (forthcoming). "Growth without governance." Economia (internet journal). _____ and Pablo Zoído-Lobatón (2002). "Governance Matters II ­ Updated Indicators for 2000/01." World Bank Policy Research Department Working Paper No. 2772. Kaufmann, Daniel and Shang-Jin Wei (1999). "Does `Grease Money' Speed up the Wheels of Commerce?" National Bureau of Economic Research Working Paper 7093 (April). Keefer, Philip (2002). Clientelism, Credibility and Democracy. Mimeo, Development Research Group, The World Bank (November). Keefer, Philip and Stephen Knack (2002a). "Polarization, Politics and Property Rights: Links between Inequality and Growth." Public Choice 111:1-2, 127-54 (March). _____ (2002b). "Boondoggles and Expropriation: When are Property Rights Secure and Public Investment Growth-Promoting?" World Bank Policy Research Report 2910 (October). Keefer, Philip and David Stasavage (2002). "Checks and Balances, Private Information, and the Credibility of Monetary Commitments." International Organization 56:4, 751-774 (Autumn). _____ (2003). "The Limits of Delegation: Veto Players, Central Bank Independence and the Credibility of Monetary Policy." Mimeo, The World Bank. Knack, Stephen (2002). "Social Capital, Growth and Poverty: A Survey of Cross-Country Evidence." In C. Grootaert and T. van Bastelaer, eds., The Role of Social Capital in Development: An Empirical Assessment, New York: Cambridge University Press. _____ and Philip Keefer (1995). "Institutions and Economic Performance: Cross-Country Tests Using Alternative Institutional Measures." Economics and Politics 7:3, 207-228 (November). _____ (1997a) "Does Inequality Harm Growth Only in Democracies? A Replication and Extension." American Journal of Political Science 41:1, 323-332 (January). _____ (1997b) "Does Social Capital have an Economic Payoff? A Cross-country Investigation." Quarterly Journal of Economics 112:1251-88 (November). Krishna, Anirudh (no date). "Seizing the Middle Ground: New Political Entrepreneurs in India." Mimeo, Sanford Institute for Public Policy, Duke University. Lederman, Daniel, Norman Loayza, and Rodrigo Reis Soares (2001). "Accountability and Corruption: Political Institutions Matter." World Bank Policy Research Working Paper 2708 (November). Leff, Nathaniel H. (1964). "Economic Development Through Bureaucratic Corruption." The American Behavior Scientist 8:2, 8-14 (November). Li, Hongyi, Lixin Colin Xu and Heng-Fu Xou (2000). "Corruption, Income Distribution and Growth." Economics and Politics 12:2, 155-82 (July). 47 Libecap, Gary D., ed. (1996). Reinventing government and the problem of bureaucracy. Greenwich, Conn.: JAI Press. McCubbins, Mathew, Roger Noll and Barry Weingast (1987). "Administrative Procedures as Instruments of Political Control." Journal of Law, Economics and Organization 3:2 (243-77). Medina, Luis Fernando and Susan Stokes (2002). "Clientelism as Political Monopoly." Mimeo, University of Chicago, Department of Political Science. Mo, Pak Hung (2001). "Corruption and Economic Growth." Journal of Comparative Economics 29:1, 66-79 (March). North, Douglass (1981). Structure and Change in Economic History. New York: W.W. Norton. _____ (1990). Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. North, Douglass and Barry Weingast (1989). "Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England." Journal of Economic History 49:4, 803 ­ 32 (December). O'Rourke, Kevin H. (1999). "Culture, politics and innovation: Creamery diffusion in late 19th century Denmark and Ireland." Mimeo, Department of Economics, University College Dublin. Persson, Torsten and Guido Tabellini (2000). Political Economics: Explaining Public Policy. Cambridge, MA: The MIT Press. _____ and Francesco Trebbi (2002). "Electoral Rules and Corruption." Mimeo, Institute for International Economic Studies, Stockholm University. Przeworski, Adam and Fernando Limongi (1993). "Political Regimes and Economic Growth." The Journal of Economic Perspectives 7:3, 51-69 (Summer). Przeworski, Adam, Michael Alvarez, José Antonio Cheibub, and Fernando Limongi (2000). Democracy and Development: Political Institutions and Well-Being in the World, 1950- 1990. New York: Cambridge University Press. Quinn, Dennis and John Woolley (2001). "Democracy and National Economic Performance: The Preference for Stability." American Journal of Political Science 45 (July): 634-657. Rauch, James and Peter Evans (2000). "Bureaucratic Structure and Bureaucratic Performance in Less Developed Countries." Journal of Public Economics 75:1, 49-71 (January). Remmer, Karen (1986). "Exclusionary Democracy." Studies in Comparative International Development 20:4, 64-68. Robinson, James A. and Thierry Verdier (2002). "The Political Economy of Clientelism." Center for Economic Policy Research Discussion Paper No. 3205 (February). 48 Rodrik, Dani, Arvind Subramanian, and Francesco Trebbi (2002). "Institutions Rule: The Primacy of Institutions over Geography and Integration in Economic Development." National Bureau of Economic Research Working Paper 9305 (November). Scott, James C. (1972). "Patron-Client Politics and Political Change in Southeast Asia." American Political Science Review 66: 1, 91 ­ 113 (March). Scully, Gerald W. (1988). "The Institutional Framework and Economic Development." Journal of Political Economy 96:31, 652 ­662. de Soto, Hernando (1989). The Other Path. New York: Harper and Row. Stasavage, David (2003). Public Debt and the Birth of the Democratic State: France and Great Britain 1688-1789. New York: Cambridge University Press. Svensson, Jakob (1998). "Investment, Property Rights and Political Instability: Theory and Evidence." European Economic Review 42:7, 1317-41 (July). Wallsten, Scott (2002). "Of Carts & Horses: Regulation and Privatization in Telecommunications." World Bank Policy Research Working Paper (October). Weingast, Barry and Mark Moran 1983. "Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission." Journal of Political Economy 91 (October): 765-800.