The Political Economy of Public Expenditures (Background Paper for WDR 2004: Making Services Work for Poor People) Philip Keefer and Stuti Khemani June 12, 2003 Governments have a crucial role to play in accelerating ­ or obstructing ­ economic development, and one of the main tools through which this role is exercised is public expenditure. Targeted appropriately, public expenditures can overcome market failures that deny large segments of the population access to economic opportunities ­ market failures such as the inability of the poor to borrow for education, or the lack of information about preventive health care, or the externalities that exacerbate public health hazards. However, distortion and misallocation of public expenditure is common, reducing the extent to which public expenditures can play a positive role in the development process. It is a common observation across countries, rich and poor alike, that substantial public expenditures are systematically misallocated, for example to wage bills for bulky state administrations, to farm subsidies that impose distortionary costs on the economy and fail to benefit the poor, and to large infrastructure projects that allow political rent extraction without creating sustainable assets, all at the expense of quality public services. These misallocations have a disproportionate impact on the poor, who are known to benefit from increased access to public services. Misallocation has persisted despite a sea change in the way in which governments are selected and remain in office. From 1990 to 2000, the number of countries governed by officials elected in competitive elections rose from 60 to 100.1 Democratization might be expected to benefit most the "median" or average voter, who in most developing countries is the "poor". Yet, public policy in emerging democracies does not seem to have tracked the preferences of poor voters. Why do policy-makers that depend upon political support from the poor not effectively deliver basic services to the poor? There are numerous imperfections in political markets that help to explain this puzzle. In this paper, we review the theory and evidence on the impact of political market imperfections, and develop the implications of these findings for the structure and design of policy interventions meant to improve the allocation of resources. Political market failures in our analysis is reduced to three broad features of electoral competition--one, lack of information amongst voters about politician performance; two, social and ideological fragmentation amongst voters that leads to identity based voting and lower weight placed on the quality of public services; and three, lack of credibility of political promises to citizens. Informed voting is costly, and voters may have difficulty in coordinating information to reward (or punish) particular politicians or political parties for specific actions that improve (or worsen) the quality of public services. Similarly, socially and/or ideologically fragmented societies are less able to provide the incentives to their 1According to the number of countries reported in the Database of Political Institutions as having competitive elections for executive and legislative office (EIEC and LIEC equal to seven). 1 political agents to improve public services, because voting is more likely to occur along the dimension of identity and regardless of policies followed by elected representatives. Even if voters are informed and coordinated in focusing on specific policies, if political competitors cannot make credible promises prior to elections, incumbents are more secure from challenge and have fewer incentives to be responsive to citizens. If politicians are credible only to a few voters, with whom they can maintain clientelist relations, then public resources are allocated to targeted benefits for these "clients", instead of to broad public services. These political market imperfections are the subject of the first part of this paper. The formal rules of the game for political markets also affect politician incentives to be responsive to the public interest. Electoral rules dictate whether politicians will campaign for the support of narrow or broad groups of voters; political institutions determine whether individual politicians are strong or weak in the making of public policy, and therefore whether the constituencies that they are more or less likely to receive services from government. The effect of formal institutions on policy outcomes is the subject of the second part of the paper. A strong conclusion of the analysis here, that is difficult for development policy, is that the most adverse effects of political market imperfections are felt in the area of broad social services, such as health and education. It is especially difficult for voters to assess the quality and efficiency of service provision and to evaluate the responsibility of specific political actors for service breakdowns. By the same token, political competitors find it especially difficult to make credible promises about service provision. Voters cannot easily collect information that would verify that politicians have fulfilled their promises. Moreover, even if they could, politicians in many countries can only make credible promises to narrow groups of voters. For these voters, it may be politically more efficient to promise narrow targetable goods, such as infrastructure provision, than it is to promise improvements in broad public services. Social fragmentation in the electorate exacerbates these problems of voter coordination in determining reward and punishment based upon political actions towards the quality of public services. To the extent that in developing countries, poor voters are more likely to vote in uninformed ways, being susceptible to campaign slogans, or polarized along non-economic ideological dimensions such as religion or ethnic identity, and political promises are particularly lacking in credibility or prone to clientelism, it is precisely the broad social services that are most likely to suffer. Evidence that social service provision has a low political priority There is a wealth of evidence demonstrating that governments prefer spending on job programs or infrastructure relative to improvements in broad social services, but that this preference is more pronounced in many developing countries. The arguments in this paper are meant to illuminate some underlying reasons for these expenditure patterns, which have not yet been as clearly identified. Feyzioglu and others (1998) find that when foreign aid is targeted at education, it is more likely to be offset by reductions in own-government financing of education. When aid is used to support public investment (transportation and communications), however, the opposite is true: governments are likely to maintain their own fiscal effort in the provision of these goods. Foster and Rosenzweig (2001) and Pande (2003) have shown that when disadvantaged groups in India ­ the landless poor and scheduled castes (respectively) ­ were newly able to elect their own representatives, government policies towards these disadvantaged groups improved. However, the 2 improvements were confined to the access of these groups to government jobs and targeted welfare transfers. Enhanced political rights did not translate into improved education or access to other broad-based social services. This kind of evidence underscores a general point that misallocation of public expenditures should not necessarily be interpreted as a result of "capture" by the "rich"--even when government expenditure programs are "pro- poor" in that they are explicitly meant to benefit the poor, they frequently take the form of targeted (and often unsustainable) redistribution programs, such as free food and temporary employment in public works, rather than broad programs to improve basic health and education services.2 Even apparently broad-based subsidies end up benefiting narrow interests. As is suggested in the case of India and as Bates (1981) has shown for Africa, agricultural subsidies, such as price protection and subsidized electricity, disproportionately benefit middle and large farmers, who grow the protected crops and are more likely to engage in capital-intensive, power-driven private irrigation. Effectively, if not de jure, such subsidies are targeted. The preference of politicians for "pork barrel" or targetable spending is not the exclusive province of developing countries, and is widely documented in the United States, for example (Mayhew 1974). The problem for development is that many governments have exaggerated preferences for targeted expenditures at the expense of untargeted health and education services that would more heavily benefit poorer landowners and landless laborers. For example, the bulk of expenditures in health and education typically flow to the salaries of teachers and health workers, yet rampant absenteeism and shirking by these service providers means that no services are effectively provided in many cases. That is, governments use these resources to provide (targetable) jobs rather than (less targetable) high quality services.3 Service delivery falls far below the levels that even poorer countries can afford, suggesting that the tradeoffs between targeted and non-targeted public expenditures are much steeper in these countries than in developed countries. Understanding the variation in political market imperfections across countries that give rise to these exaggerated preferences is the goal of this paper. In particular, the paper asks: why are voters not able to provide stronger incentives for politicians and service providers to deliver better outcomes? Demand side considerations in poor public service delivery Although the analysis here focuses on the supply side ­ the incentives of political actors to provide public services ­ one might naturally be concerned that the political differences among countries are spurious, and that the problem is rooted in the demand side. 2 In their review of research on public spending and the poor, Van de Walle and Nead (1995) find that resources spent on basic health and education services ("broadly" targeted fiscal policies) have higher payoffs for the poor than finely targeted food subsidies or other redistribution schemes, in part because of the administrative costs and behavioral consequences of the targeted schemes. 3 InIndia, recurrent expenditures on primary education accounts for 98 percent of total government expenditure on primary education; salaries account for 96 percent of recurrent expenditures, and teachers' salaries account for 97 percent of all salaries in education spending (Tilak, 1993, p. 60). Yet, field investigations in rural areas of Indian states, particularly in the north, reveal that teacher absenteeism is endemic, with almost two-thirds of the teachers employed in the sample schools absent at the time of the investigators' unannounced visits (Drčze and Gazdar, 1996; Weiner, 1991; Prasad, 1987). 3 Voters in poor countries may simply not value health or education provision by the government. Scheduled caste members and the landless poor may similarly be unaware of the importance of education or, aware of the importance of education, but nevertheless prefer that their politicians provide other services. Ignorance of the value of education seems not to be a sustainable thesis. A common finding of village studies and household surveys in India is that education is widely perceived by members of disadvantaged groups as the most promising chance for a better life for their children (Drčze and Sen, 1995). Farmers, landless laborers, and scheduled caste members therefore know the importance of education for their children. A thought experiment also makes clear that demand side considerations alone cannot explain why, though fully informed about the value of education, the landless poor might still prefer targeted jobs. Suppose that a village must choose between demanding a single job for some, randomly chosen individual in the village, or the presence of a teacher from outside the village. Suppose the village has thirty families and that the teacher would instruct one child from each family. Further, assume that literacy raises the present value of each child's lifetime income by 30 percent, that literacy takes five years of education to achieve, and that the present value of each child's income, absent education, would be the same on average as that of the average wage earner in any of their families. Then the total value to all 30 families from having the teacher (assuming the teacher is guaranteed to stay for five years) would be 0.3 * 30 * (the present value of average wage earner's lifetime income). To match this wealth effect, and even assuming that the job offered in lieu of a teacher was guaranteed for the life of the recipient rather than for only five years, the job would still have to pay ten times the average wage. The demand for education would have to be extraordinarily low, either because of high discount rates or severe credit market constraints that block access to complementary inputs (books, foregone child labor), to explain the apparent preference of the poor in many developing countries for jobs and subsidies over schools and clinics. However, because we observe the poor making large sacrifices to educate their children ­ for example, in private institutions ­ despite imperfect credit markets and discount rates that are not particularly low, it makes sense to probe further and investigate why voters in some countries do not pressure politicians to provide high quality education. Part 1: When does political competition lead to optimal social services? The Benchmark Case A simple illustration indicates the stringency of the conditions under which elected officials always provide the socially optimal level of public goods. By the same token, it clarifies the many sources of imperfection in political markets that can undermine the incentives of political decision makers to provide optimal government services to citizens. We can imagine a country in which all political promises are credible, politicians cannot make targeted transfers (transfers to some voters but not others), voters are identical, and voters can observe at no cost the contribution that politicians make to their welfare. Under these conditions, the public good preferences of voters always translate into actual government policy. To see this, assume there are identical voters, whose individual welfare W is given by c + V(g), where g is the amount spent on public goods(?) per-voter cost of public goods, consumption c isy(1 - ), or income net of a uniform tax, and g = y. 4 The welfare of the average or median voter (the same in this case, since all voters are identical) is therefore maximized by public good provision g* such that Vg(g*) = -1/y.4 This is precisely the level of public good provision that maximizes total social welfare (taken to be the sum of the welfare of all voters). Any candidate who deviates from the policy promise g* and = g*/y is defeated by any candidate who offers this optimal policy package, since all voters prefer g* to any other g. The example is unrealistic for many reasons. However, each of the deviations from reality illuminates a political obstacle to social service delivery. One key assumption of policy making in the ideal world is that pre-electoral promises are assumed to be binding on the election winners. However, electoral promises are not legally enforceable anywhere. Instead, the credibility of campaign promises depends on the reputation of either the individual candidate making the promises or of the political party to which candidates belong. However, in many cases ­ such as young democracies ­ these reputations are weak, or are developed only for a limited number of issues. For example, individual politicians or parties may be well-known and believable as freedom fighters or as defenders of a religious faith, but have no reputation with regard to education or health services. If policy promises are not credible, voters do not believe challengers who claim that they will adopt policies different than those pursued by incumbents. Given challenger ineffectiveness, voters can do no better than to establish some threshold of performance that the incumbents should meet if they are to be re-elected. This substantially weakens voter control over incumbent performance and, as Ferejohn (1974) demonstrates, allows incumbents to pursue policies substantially at odds with the ideal world. Moreover, if voters cannot coordinate even on a performance standard for the incumbent, as in Robinson and Verdier (2002), politicians never pursue the public interest except to the extent that it directly improves their own private welfare. In this limiting case, non-credible politicians set the tax rate to maximize the rents that they can extract from the economy during their expected time in office and provide no public goods whatsoever. When politician promises are entirely non-credible and voters unable to coordinate on a response to badly performing politicians, elections are almost meaningless. This is a crucial lesson that reappears throughout the discussion in this paper: any set of institutions that can sustainably guarantee accountable and responsive government includes competitive elections, but elections by themselves are far from a sufficient condition for good governance. An additional assumption in the idealized model is that voters can costlessly observe the contribution of politicians to government policy changes and the connection between government policy and their own welfare. Politicians therefore always get credit or blame for the actions they take or fail to take on behalf of voters. However, if voters cannot verify that politicians have fulfilled their promises to them, then again politician promises are not credible. The same outcomes emerge as before, in the non-credible politician case. If voters are utterly ignorant about politician responsibility for government policy or about the 4Rewrite welfare in terms of g and maximize with respect to g. 5 contribution of policy to their own welfare, the results from the utopian model entirely reverse and no public goods at all are provided. Voters are also assumed to be identical in the idealized model. Obviously, though, voters differ on a host of margins, including income, ideology, religious belief, language, the value they attach to different public goods, occupation and location. All of these affect their preferences for government policies and public goods. These differences can lead to distortions of several kinds. First, in most countries, the distribution of income is skewed to the right ­ towards the higher end of the income spectrum ­ so that the income of the median voter is less than the average income of all voters. Under these circumstances, government should be larger and social services should be correspondingly more extensive (Meltzer and Richard, 1981). However, this prediction actually exacerbates the puzzle of why many countries severely under-provide social services, since it is in many countries where the poor are the median voters that social services are often the most woeful. Moreover, there is little evidence that redistribution is greater in countries with greater income inequality (Knack and Keefer 1997). Ideological distortions have an effect in the opposite direction. If voters have strong preferences for or against specific candidates or parties, independent of the policy choices of these parties, social service provision can suffer significantly. It is easy to show the potentially devastating impact of ideological preferences on outcomes in the ideal world. Imagine that there are only two political parties competing for votes. Some voters derive utility from the election of party A and disutility from the election of party B, and the remaining voters derive utility from party A and disutility from party B. For simplicity, voters who prefer party A are identical in their preference for party A, as are the voters who prefer party B. Further, the disutility that voters get when their non-preferred party is elected is exactly equal to the utility they get if their preferred party is elected, and the amounts of utility and disutility for each group of voters are the same. That is, for all voters, the utility they get if their preferred party is elected is , and the disutility they get if their non-preferred party is elected is -. As before, politicians representing each party make promises regarding public goods and elections are held. When ideological polarization is severe enough, however, politician promises regarding public goods are irrelevant. To see this, assume that the society is completely polarized, as with countries emerging from a period of civil war or that have a history of ethnic or religious conflict, and that half of all voters prefer party A and half prefer party B . If party A is elected, voters who prefer party A receive WA = c + V(gA) + and voters who prefer party B receive WB = c + V(gA) - , where gA is the level of public goods promised by party A. Conversely, if party B is elected, welfare outcomes for each group of voters are given by WA = c + V(gB) - and voters who prefer party B receive WB = c + V(gB)+ .. If ideological preferences are zero, = 0, the two parties both promise g*. However, as grows, voters become more willing to vote for their preferred party even if that party promises fewer public goods than the non-preferred party. If social polarization is sufficiently high and sufficiently large so that V(g*) + y(1 ­ g*/y) - < V(g) + y(1 ­ g/y) + for all g, or V(g*) - V(g) + g ­ g* < 2 , there is nothing that the non-preferred party can promise prior to the election that will convince the supporters of the other party to switch 6 their allegiance. The welfare that the supporters of the other party receive from electing their own party, even if their own party provides no public goods at all, is greater than the welfare that they receive if the other party is elected and provides the optimal level of public goods.5 Finally, in the ideal world, only public goods can be provided. However, in every country the ability to make targeted transfers is a key weapon in the political arsenal and almost always directly conflicts with the goal of providing efficient public services to those who most benefit from them. Benefits are targeted according to political calculations that need not and often do not coincide with technocratic notions of targeting, which emphasize targeting according to other criteria, such as need. Political and program targeting criteria need not ­ indeed, usually do not ­ coincide, undermining adequate social delivery. To see the implications of political targeting, one can assume identical voters, as before, and allow individual voter welfare W to be given by c + V(g) + mi, where mi is a transfer from the government to voter i. As before, c is y(1 - ), but y = g + mi. For even the slightest deadweight loss from taxation, m should be zero since the utility gain of transfers to any voter is offset by an equal utility loss borne by other voters. However, there are many possible realistic scenarios under which politicians would prefer only to provide m. If voters can only verify promises related to the direct provision of targeted goods (e.g., a job), but not of broad public goods (e.g., efforts to improve school quality), politicians have an incentive to provide only transfers to voters. 6 If most voters are ideologically polarized, but some are not, politicians would compete by making promises only to the ideologically uncommitted. If this group were small enough, its members would prefer all tax revenue to be used to provide them with targeted transfers than with public goods that would benefit all citizens.7 These simple theoretical illustrations provide a conceptual framework for a more careful review of the different factors that interfere with the ability of voters to hold politicians accountable for the quantity and quality of social services. The remainder of this section reviews each of these possible imperfections in political markets, beginning with the role of information. 5 Governments cannot extract rents in the ideal world ­ if they could, it is easy to see that in a polarized world, the ability of governments to extract rents would be higher. 6 In the absence of additional structure, there is no equilibrium policy outcome in the simple model of targeted transfers. There is no offer of transfers and public goods that is invulnerable to defeat by some other offer. Transfers to each voter each constitute a different policy dimension and Plott (1967) shows that majority rule with a multi-dimensional policy space is unlikely. Many of the political institutions discussed in this paper solve this problem. For example, most political systems grant agenda setting authority to some elected officials but not to others. Those officials can make take -it-or-leave-it offers to other policy makers, thus offering an institutional solution to the problem of policy instability introduced by politically targeted transfers. Such institutional innovations tend to give rise to their own distortions. 7 That is, for any amount of tax revenue, a small group of k swing voters would prefer transfers to public good provision as long as Vg <1/k, where 1/k is the marginal utility gain for members of this group of an additional unit of transfers to them. 7 Information and the provision of social services to the poor Information problems take several forms, at the heart of which lies essentially a free- rider problem: the private costs of informed voting exceed the private benefits, but the social benefits of a single voter's information far exceed the costs of information collection. Early political studies on voting behavior approached any individual's decision to vote at all as a conundrum, since the private cost would be non-negligible while the private benefit, in terms of impact of a single vote on election outcomes, is essentially zero (Downs, 1957). Nevertheless people vote, even on the basis of limited information. There are at least four specific sources of imperfect information in political markets. First, voters may not be able to disentangle the welfare effect of exogenous events and government decisions (or lack thereof). Wolfers (2002) finds that voters in oil-producing states in the US tend to re-elect incumbent governors during global oil price rises and vote them out of office when the oil price drops. Anecdotal evidence from India suggests that governments tend to lose elections in years of poor rainfall. If voters are easily confused by "noisy signals" created by exogenous shocks, political incentives to exert effort on behalf of citizens are correspondingly reduced.8 If outcomes in public services such as health and education are more noisy and difficult to attribute to the efforts of political agents, perhaps because they are determined by several factors outside the direct control of political agents, then public resources may not be optimally allocated towards them, even if there is high social pay-off from investing in such services. Mani and Mukand (2002) show that if elections serve the purpose of voters choosing amongst candidates to select the most competent one, then resource allocation will be biased against those public goods whose outcomes are more noisy and harder to use to assess politician ability, as politicians will have the incentive to provide other goods that are better signals of high ability. Second, voters may make use of very limited information in determining electoral choices, such as behaving myopically and placing greater weight on the most recent decisions of politicians, independent of their effect on voter welfare relative to earlier decisions. This is a particular problem for investments in some public services for which the measurable outcomes of policy change may not emerge for several years after the policy action has been taken. The effects of education reforms, for example, are typically difficult to verify until a cohort of students has been exposed to them for a sufficient period of time. If information constraints drive voters to place excessive weight on the recent or the visible, they create perverse political incentives just before elections to boost spending and to divert funds to short-term visible objectives shortly before elections. Although evidence for electoral budget cycles in developed countries is at best ambiguous, Shi and Svensson (2002), Block (2002), and Schuknecht (2000) find large electoral cycles in monetary and fiscal instruments in developing countries. To the extent that voters in such countries are less well-informed and have less experience with democracy, these findings are consistent with the information explanation for such cycles. 8 Filteringout performance from noisy signals in outcomes is a pervasive issue in principal-agent problems, and not somehow restricted to or especially intractable for the citizen-politician agency problem. Bertrand and Mullainathan (2001, p. 1) find, for example, that CEO pay "responds about as much to a lucky dollar as to a general [earned] dollar". 8 In contrast, for an older democracy in the developing world, India, Khemani (2003), does not find evidence of significant changes in aggregate fiscal variables at election times in state governments, but does find that the composition of spending and revenues changes, possibly to target special interest groups in exchange for campaign support. As the time of elections draws near, state governments in India increase expenditures on public investment projects and away from more broad-based categories of spending on public services, and provide targeted tax breaks to narrow groups of producers possibly in exchange for campaign finance. This leads us into the third type of information processing problem. Rather than being influenced by independent evaluations of the performance records of competing candidates, voting decisions may be swayed by political campaigns and advertisements, thereby creating a role for special interests to purchase narrowly targeted policies by providing campaign finance (Baron, 1994, and Grossman and Helpman, 1996). Special interests could be single firms, groups of manufacturers, farmers, public sector employees, or subsidized retail store owners.9 They need not be rich, but they are, by definition, informed. To the extent that the poor are disproportionately uninformed, therefore, the organized poor are likely to be under-represented as special interests. Similarly, to the extent that voters in developing countries are less informed, broad public policies are likely to be distorted in favor of special interests. The presence of voters who are uninformed about politician attributes also provides an incentive for politicians to extract rents. Models that make these predictions assume that uninformed citizens vote randomly (probabilistically). Persson and Tabellini (2000) argue that the larger the density of citizens who vote randomly (that is, with complete disregard for policy), the greater is political rent extraction in equilibrium. Fourth, even if voters can perfectly observe the effects of government action on their welfare, they may be unsure of which particular government actors were responsible for the action and whether the actors that they personally can elect could influence the action. This again reduces the incentives of government actors to ensure good outcomes. Health and education, which are the products of decisions at every level of government, down to the individual service provider, are particularly vulnerable to this type of incomplete information. Detailed behavioral studies in the US have shown that voters by and large tend to adopt simple voting criteria based on very limited information about politics and public policies (various articles in Ferejohn and Kuklinski, 1990). However, Fiorina and Shepsle (1990) and Chappell and Keech (1990) argue that citizens can employ voting rules requiring very little information and still motivate politicians to pursue policies in their interest. Ferejohn (1990, pp 8-9) paraphrases this process as follows: "find a way to get the electorate to commit itself to act as though it is a simple principal with a one-dimensional set of rewards. In this way, incumbents will be prevented from taking advantage of the conflicting interests in the electorate". Hence, if the electorate in aggregate is able to coordinate on a common set of issues then politicians would have incentives to pursue policies as defined by those issues. 9 Alderman (1988) describes protests in Pakistan by retailers of government subsidized wheat and flour when their livelihood was placed at risk because of amendments in the subsidies program. 9 Yet, information constraints in developing countries undermine this optimistic view of voter behavior, and particularly so for providing the right incentives for the provision of broad social services. Fiorina (1990) emphasizes that most of the information people use to make voting decisions is essentially "free" in that it comes with the ordinary performance of social and economic roles. But the content of "free" information can vary widely over the electorate depending on the differentiation of occupations in the economy and the variability of social settings. One result is that citizens will specialize in information about some things rather than others (Ferejohn, 1990; Iyengar, 1990; Ottati and Wyer, 1990). It is a small leap to the conclusion that the information base of poor, rural citizens of developing countries is therefore skewed in a way that detracts from their ability to hold elected officials accountable for the quality of public services. Poor voters have limited access to universal sources of information, such as newspapers and radio, that can help coordinate information about broad social services. They are less likely to be able to jointly evaluate the quality of services such as health and education based on their individual experience, as these services are transactions-intensive and depend critically on day-to-day provider behavior, and outcomes are sensitive to provider discretion. Furthermore, the efforts of remote political representatives to improve actual quality of these services may be harder to be jointly informed about than simple observations of whether they built new schools and clinics, and provided salaries for teachers and doctors. Alternate programs of targeted transfers that provide private benefits such as subsidized food, and jobs in the public sector, may be easier for poor voters to coordinate upon in determining re-election of incumbents or in choosing between rival candidates. Information problems may thus lead voters to give greater credit to politicians for initiating public works projects (including school construction), providing direct subsidies for essential commodities, and increasing employment in the public sector (including hiring teachers and doctors), than for allocating resources and effort towards improving actual quality of education and health services, such as ensuring teacher and doctor attendance or that school and clinic buildings are properly equipped. There is substantial evidence that voters with access to information are more likely to receive transfers. Stromberg (2001) finds that between 1933 and 1935 in the US, federal assistance to low-income households was greater in those counties where more households had radios and were thus more likely to be informed about government policies and programs. The spread of the radio particularly improved information access for rural voters, who were previously disadvantaged relative to urban voters (since the latter already had access to alternative sources of information such as newspapers, while radio waves made it easier to deliver information to remote areas). It accounted for as much as 20 percent greater allocation of social assistance funds to a rural county as compared to an identical urban county. Besley and Burgess (2003) find that state governments in India are more responsive to declines in food production and crop flood damage via public food distribution and calamity relief expenditure when newspaper circulation, particularly in local languages, is greater. Although these cases show that informed voters have more leverage, they do not show that as voter information increases, public good provision improves. It could easily be the case that the mass media better enabled politicians to take credit for targeted payoffs to particular constituencies, leading them to reduce expenditures on public goods to better take 10 advantage of this opportunity. However, more optimistically, these transfers could also have been financed by reduced rent-seeking on the part of politicians. The importance of mass media in informing voters is similarly illustrated by Shi and Svensson (2002) who find that some of the difference in the size of political budget cycles across countries is due to variation in access to free media ­ the greater is access, the smaller are the observed budget cycles. To the extent that political budget cycles are biased towards the provision of targeted goods to voters, their conclusions support the argument that politicians provide more broad public goods when voters are better informed. Social polarization and provision of services to the poor Politicians can under-provide public goods when voters are uninformed, but also when voters select candidates on the basis of social and ethnic identity, independent of policy performance or promises. Both information constraints and social polarization lead votes to be cast in theoretically identical ways that give little weight to the quality of public services, even though voters may have perfectly rational and high demand for such services. Studies of electoral politics in India show that identity characteristics along ethnic, linguistic and religious lines dominate political behavior (Weiner and Field, 1974). Similarly, in Nigeria, social cleavages along the lines of religion and ethnicity play a prominent role in determining political behavior. In socially polarized and/or ethnically fragmented societies, voters therefore tend to vote for those candidates they most closely identify with, irrespective of public performance and policy records; political competition between parties thus also concentrates on identity issues, and candidates are nominated from constituencies largely on the basis of demographic calculations of ethnicity and religion. These voting patterns are not the same as "ideological" voting in the US or Western Europe. Surveys of voting behavior in the US in the 1950s and 1960s (Campbell et al, 1966) show that an overwhelming determinant of voting behavior is the ideological identification of voters with political parties. In this case, though, ideology has significant policy roots ­ in citizen identification with the perceived position of political parties on a wide range of social and economic issues. Why the difference between India and Nigeria, on the one hand, and the US on the other? There are two possibilities. One, which is the implicit argument in much of the literature, is simply that individuals in some countries simply have a strong preference for members of the same ethnic, religious or social group. As a consequence, the utility they derive from providing public goods to all is lower than it otherwise would be. A second possible explanation receives more attention in the next section of the paper: where political competitors are generally not credible, voters rely on personal connections to competitors to establish which are credible and likely to provide them benefits once in office. In the US, parties have established credible positions on certain issues, which together comprise a package that attracts particular voters who then appear to be "ideologically" committed to the party ­ when in fact, they are committed to the policy positions with which those parties are credibly associated. In India and Nigeria, though, the problem is the reverse. Voters only believe promises from political candidates within their own ethnic or religious groups; those promises are therefore, necessarily, narrow and targeted to members of the respective ethnic group. Political promises for broad-based public services that cut across social groups are not credible. 11 The evidence is consistent with either explanation. In a cross-country setting, Easterly and Levine (1997) find that ethnic diversity is negatively correlated with the provision of public goods such as the percentage of roads that are paved, efficiency of the electricity network, and years of schooling of the population. For cities in the US, Alesina, Baqir and Easterly (1999) show that shares of public spending on productive public goods ­ education, roads, sewers, and trash pick-up ­ are inversely related to the city's ethnic fragmentation. Miguel (2001) concludes that higher levels of local ethnic diversity in Kenya are associated with sharply lower contributions to primary school funding and worse school facilities. Animus across ethnic groups explain these outcomes directly. However, if individual politicians can only make credible promises to members of their own ethnic groups, as a later section explains, it makes little sense for them to provide public goods that benefit all ethnic groups. This also explains these findings. Where identity matters, minority identities (minority castes or religions) are likely to suffer in public policy decisions decided by the majority. In India, Betancourt and Gleason (2000) find that districts with a higher proportion of traditionally disadvantaged groups with respect to caste and religion ­ the scheduled castes and Muslims ­ have lower public inputs in health and education. Similarly, Banerjee and Somanathan (2001) find that districts with a higher share of scheduled tribes in population receive significantly fewer "desirable" public goods. They also find some evidence that links the relationship between ethnic heterogeneity and public good delivery to underlying political incentives--districts that are ethnically fragmented are also likely to be politically fragmented, in that elections in these districts are characterized by a larger number of contestants and a smaller vote share for the winning party. There is also substantial anthropological and anecdotal evidence in India that disadvantaged groups are systematically excluded from using public goods within their own villages by social processes of discrimination. Micro-level case studies and survey evidence from India show that within-village inequality in education access and achievement is significant, with the privileged castes in the village enjoying near-universal adult literacy for several decades while literacy rates are still close to zero among disadvantaged castes in the same village (Drčze and Sen, 1996). To solve the problem of animus and social discrimination, India instituted political reservations for scheduled castes and tribes in the national and state legislative assemblies. Pande (2002) finds that mandated reservations in state legislatures resulted in increased public sector job quotas for scheduled castes in the public sector, but significantly lower resources allocated to education. The argument here suggests that this is because narrow appeals to ethnic constituencies are the only credible promises that politicians can make in many ethnically fragmented societies, leading them to favor targeted rather than broad public goods. Mandated reservations may in fact worsen overall public service performance by strengthening clientelist relations and reducing incentives of political competitors to invest in broad policy reputations across the electorate. Unfortunately, though parties could gain by attempting to build such reputations, there are also strong political incentives in the opposite direction. As Glaeser and Shleifer (2002) have shown in the case of the city of Boston, as the experience of Zimbabwe makes self-evident, and has been the case throughout history, politicians can often gain political advantage by fanning ethnic divisions. In the limit, as in Boston or Zimbabwe, electoral 12 advantage is cemented by simply pushing the out-of-power ethnic group out of the jurisdiction. Credibility and the provision of social services Credibility influences the provision of social services in two important ways. First, credibility sharpens the effects of competitive elections on politician incentives. When campaign promises are not credible ­ when it costs election winners little to abandon them ­ electoral competition has relatively modest effects on service provision. Public goods are likely to be under-provided in less credible political environments. One aspect of this has already been discussed: when the quality of social services is difficult for voters to observe, voters cannot easily verify whether quality public goods have been provided or not. Absent this verification, politicians have an incentive to under-provide relative to what they have promised. Second, credibility affects the types of public goods that politicians offer. In countries where voters believe political leaders are less secure in their positions, they place less value on promises by political competitors to improve the quality of education, which are of no value unless they are carried out for a substantial period of time, and more value on promises whose benefits are immediately realized. These two factors are discussed in greater detail in the sections that follow. The credibility of political promises and the impact of clientelism on social service delivery If pre-electoral promises by competing parties and candidates are not credible, voters lose leverage over the actions of political actors. For example, refining a model developed by Ferejohn (1974), Persson and Tabellini (Chapter 8, 2000) demonstrate that public good provision falls substantially below the socially optimal amount when politicians cannot make credible promises. Like Ferejohn, they assume that voters can coordinate on a minimum threshold of incumbent performance such that failure to meet the threshold results in the incumbent's expulsion. However, such standards are difficult to formulate: if they are too high, politicians no longer value re-election, or may believe that it is impossible to get re- elected, and so reject any social service provision in favor of an aggressive effort to accumulate resources for themselves. If standards are too low, politicians have no need to provide high quality social services in order to be re-elected. When standards are exactly at the optimal level, just high enough such that incumbents still have an incentive to meet them and get re-elected, social service delivery is worse than if all assurances prior to choosing government leaders were credible. Under these circumstances and if, for example, politicians need to attract the votes of 51 percent of the electorate in order to be elected, incumbents provide public goods as if only 51 percent of voters benefited from them rather than all voters. This implies significant under-provision of the public good. An attempt by any majority of voters to demand a higher level of public goods would be undercut, because the incumbent could always offer targeted transfers to a few voters that offer greater welfare than the extra public good but are cheaper for the incumbent to provide. However, since all voters would like to receive such transfers, they bid them down to zero. In other models (Robinson and Verdier 2002, for example), even this level of public good provision is not possible: not only can political competitors make no credible 13 promises, but voters cannot coordinate on the threshold that they will use to judge incumbents. Incumbents therefore approve only government expenditures from which they themselves personally benefit, such as roads that pass by their privately-owned factories. In the context of the Ferejohn/Persson and Tabellini models, where government actors benefit materially only from rents, the Robinson and Verdier electoral scenario implies no public good provision whatsoever. It is rarely the case that no politician is credible to any voter. However, in many countries the foundation of a politician's credibility is not based on the politician's policy record or history of policy stances, bolstered perhaps by the policy record of the political party to which the politician belongs. Instead, some voters believe some politicians who have shown themselves to be reliable sources of personal assistance. These might be locally influential people who have helped families with loans or jobs or assistance with legal or bureaucratic difficulties. In the absence of well-developed political parties or national party leaders who are more broadly credible to voters, the promises of such influential people are all that voters can rely on in making electoral choices. Many observers of political competition and political decision making in developing countries have argued that it is characterized by an excessive tendency of political patrons to provide private goods to clients. This excess is called clientelism. Keefer (2002b) shows that clientelism can be viewed as the natural outcome of political competition when the credibility of political competitors is limited. In these cases, political promises are credible only to "clients." This has precise implications for policy: the larger the number of clients, the greater the focus of government spending on items targeted to specific individuals (clients) and, ultimately, the less spent on public goods. Compared to a situation where no politicians are credible, clientelism generates less rent-seeking or corruption ­ but only because instead of keeping resources for themselves, patrons are obligated to transfer the fruits of office to clients. When only clientelist promises are credible, promises of buildings and government jobs become the currency of political competition at the expense of universal access to high quality education and health care. The former can be targeted to individuals and small groups of voters (clients) and therefore are clearly seen as evidence of political patrons fulfilling their promises to clients. Universal access to high quality services is by definition not easily targeted. Even quality, alone, is difficult to target. Voters are less likely to observe the contribution of political effort when they observe quality improvements in teaching than when they observe a teacher in a school that previously had none. Voters cannot distinguish whether the quality or attendance of teachers has improved because of their own pressure on the teacher, because the teachers themselves decided to do a better job, because of a generalized reform in teacher quality that is affecting all teachers and education beneficiaries, or because of the targeted intervention of a particular politician. There is no evidence on the effect of credibility and clientelism on social service provision, but there is indirect evidence supporting the notion that clientelist governments have a stronger than average preference for targeted infrastructure provision and tend to be more corrupt than average. Keefer (2000b) argues that the age of democracy is one important characteristic of countries that influences whether political competitors are non- credible or clientelist or not. Because they are young, with fewer years and elections with which to have built up policy reputations, political competitors are less likely to be able to make credible promises to all voters and are more likely to rely on clientelist promises 14 (targeted promises to specific individuals or groups to whom they can make credible promises for various reasons, including past dealings with them). As young democracies age, politicians are likely to increase their credibility first on a targeted basis ­ they are first likely to focus on increasing the number of clients they have. Keefer (2002b) summarizes evidence showing that, in fact, targeted spending ­ public investment ­ is higher in young democracies than in old. Moreover, as young democracies age and the number of clients rises, targeted spending in the form of public investment increases. Keefer (2002b) also reports that corruption falls as democracies age, and that corruption is significantly higher in young than in established democracies. These results are relevant for social service delivery since social services are likely to be of lower quantity when public investment is high and of lower quality when corruption is high. Qualitative evidence directly bearing on education supports the statistical evidence on property rights. For example, competitive elections were a regular feature of the political landscape of Pakistan during the 1990s, but neither political parties nor political leaders could make credible promises to voters. Reliance on clientelist promises in pre-electoral competition was therefore pervasive. One would expect the provision of broad-based public goods in such an environment to be low and the provision of targeted goods to be emphasized. Moreover, for decades the country has been marked by considerable political instability; specifically, the tenure of incumbent leaders was frequently cut short, as happened three times in the 1990s. This should generate a marked preference for the provision by government of goods that generate benefits immediately and are not dependent on the decisions of future governments. In fact, compared to countries with similar incomes per capita and demographic characteristics (age and proportion rural), access to potable water in Pakistan in 2000 was 25 percentage points higher than expected. Investments in potable water, particularly in rural areas where they consist largely of well-drilling, are particularly easy to target and their benefits are immediately accessible and observable. Primary school enrollment, though, was 20 percentage points less than one would have expected, controlling for the same variables (World Bank 2002). Recent instances of reform show that vigorous action by national leaders can short- circuit clientelist pressures, though possibly in exceptional circumstances. President Museveni of Uganda eventually made his commitment to education public and emphatic. His specific reforms were visible and easily observable by voters, however: increasing the fraction of capitation grants to schools that schools actually received. On the one hand, therefore, entrepreneurial politicians in clientelist environments can extract significant political benefits from non-targeted policy improvements by making investments in communication that firmly link their reputations to the realization of their policy promises. On the other hand, however, such policy improvements have to be measurable and observable by voters. Hence, Museveni's emphasis on capitation grants and local school financing, rather than on curricula and teacher quality. Horizon problems and credibility A second credibility problem emerges when political competitors, whose term in office is expected to be short, make promises that are credible for as long as they are in office. This leaves them unable to credibly promise to implement projects that require a longer time period to bear fruit. This matters little if the promises concern jobs or public works projects, which, in principle, can be fulfilled with little delay after an election and 15 whose benefits can be quickly realized and recognized by voters. However, shortened political horizons render other government services, such as education, considerably less useful. Education must be received for a certain minimum period of time before recipients realize any benefits from them. This is particularly true in the early years: schooling yields few benefits until students achieve literacy and numeracy, but this may not occur before the fourth grade. Using data from Spain, Sanmartín (2001) finds generally low returns to education prior to 10th grade, but a spike in returns to education upon completion of the fourth grade. Glewwe and Jacoby, in examining the determinants of student achievement in Ghana, gave subjects a basic reading and mathematics test. Performance by those who had not advanced beyond grade school was so poor that they were excluded from the sample and the authors focused on middle school achievers only. Educational credentials also serve a signaling function that exacerbates this horizon problem: receipt of the credential generates a discrete jump in the returns of the previous years of schooling, but the credential can only be received after a certain number of years of schooling. The magnitude of the signaling effect is contentious, particularly for primary and secondary education. Layard and Psacharopoulos (1974) strongly dispute it, but Hungerford and Solon (1987) claim to find convincing evidence of it. The horizon problem can have a significant impact on voter decision making. Recalling the earlier example of 30 families, each with a school age child. They can choose between a politician who promises to provide a job and a politician who promises to staff the empty school with a qualified teacher from outside of the 30 families. Prior to the election, they do not know which family will get the job, so the expected value to each family of the job is 1/30 of the salary. If the teacher is certain to stay for five years, each child becomes numerate and literate, raising expected lifetime income. Otherwise, there is no benefit to education. If politician promises are credible over five years, the voters prefer the teacher; otherwise, they do not. Again, indirect evidence suggests political horizons affect government policies with long-term effects, like education. Clague, et al. (1996) consider the security of property rights, another outcome of state institutions that is sensitive to the horizons over which government actors can make credible promises. They make the argument that the older is a regime (the years a country is democratic or that an autocrat is in power), the longer is the horizon of the political decision makers. The evidence they present suggests that the security of property rights is higher in both older democracies and older dictatorships. Keefer and Knack (2002) find as well that productive public investment ­ the payoffs to which are largely in the future--fall when property rights are weak, while unproductive public investment ­ corruption and rent-seeking ­ rise. These results are suggestive, at least, that in the area of social service provision, particularly education is likely to suffer when the horizons over which government officials can make credible promises are short. Information and credibility Reforms affecting teacher performance in the classroom or rates of infection in hospitals are not easily observed by voters. This lack of observability discourages political actors from moving on such reforms: even if they promise higher quality services, voters cannot easily verify that service quality has improved. For example, voters can easily verify that politicians have complied with a promise to bring a teacher to their school. They can 16 less easily verify that the teacher is good, or that the teacher will remain for a sufficient period of time to deliver high benefits to the village. If politicians cannot take credit for their efforts to improve teacher quality, they provide and voters expect low-quality teachers. The quality problem is exacerbated when the political rewards are high from targeting particular providers for contracts or jobs, including teaching jobs. Since voters do not give the politician credit for providing them with high quality teachers, anyway (since they cannot observe quality), politicians have an incentive to use their clout to fill teaching jobs with the nieces and nephews of constituents, regardless of their professional qualifications. In fact, the lower the quality of the service provider, the higher the rents that providers receive and the greater their debt to the politicians. Gazdar (2000) and the World Bank (1998, 2001) make it clear that non-professional qualifications are key factors in the placement of teachers in Pakistan. Similar reports from the Dominican Republic suggest that posting of teachers is highly discretionary and not clearly related to educational concerns (Keefer 2002a). The dynamics of political competition and the provision of social services: an example from India Some of the most striking contrasts in basic health and education outcomes exist between neighboring countries with comparable levels of economic development, and between regions within the same country ­ between Sri Lanka, Thailand, and the southern states of India on the one hand, and Pakistan, Bangladesh, and the northern states of India on the other. One such contrast, between the northern state of Uttar Pradesh (UP) and the southern state of Kerala in India, has been explored in depth by Drčze and Sen (1995, 1996). These two states exhibit almost identical levels of per capita income and poverty, but dramatically different outcomes in health and education. Drčze and Sen attribute these stark differences to the divergent nature of public action in these respective states. We reinterpret their example to emphasize the differences in political incentives of policy-makers across the two states to provide social services to all. Table 1 (based on data presented in Drčze and Sen, 1996) summarizes this contrast. Human development outcomes in Kerala are comparable to some of the richest nations of the world, while those in UP are similar to those in the poorest nations of the world, despite the two having very similar economies, levels of per capita income, and poverty. Figure 1 suggests that the difference in human development outcomes is due, at least in part, to the striking differences in real per capita public expenditures on health and education in the two states. Over four decades from the 1960s to the 1990s, average real per capita spending in each decade in Kerala has been more than double that in UP. Figures 2a and 2b show spending on health and education and spending on state administration alone (on the overall organs of the state, interest payments, pensions, etc., that is, exclusive of spending on public services) as a proportion of total expenditures in the two states since 1960 to 1998. Kerala started out in the early decades of electoral competition investing more than three times the proportion of its public resources in social services compared to UP. Largely because of the growing debt burden in both states which increased the share of resources going towards interest payments (included under state administrative spending), they have converged to similar levels of spending, in proportional terms, in recent years. 17 Table 1: Contrasting Experience in Social Services India Uttar Kerala Pradesh (UP) Per capita income at current prices 1991-92 (in Indian rupees) 5583 4012 4618 Poverty headcount ratio, 1987-88 (%) 44.9 47.7 44 Real (1992 Rs) per capita public spending on education, 1985-92 228 147 309 Real (1992 Rs) per capita public spending on health, 1985-92 70 49 82 Literacy rate, age 7+, 1991 (%) Female 39.3 25.3 86.2 Male 64.1 55.7 93.6 Percentage of rural children aged 12-14 who have never been enrolled in a school, 1986-87 Female 51 68 1.8 Male 26 27 0.4 Proportion of persons aged 6 and above who have completed primary education, 1992-93 Female 28.1 21.4 60.5 Male 48.6 47.3 65.8 Proportion of children aged 12-23 months who have not received 30 43 11 any vaccination, 1992-93 Proportion of recent births preceded by an antenatal check-up, 49 30 97 1992-93 Infant mortality rate per 1000 live births, 1990-92 80 98 17 Proportion of villages with medical facilities, 1981 14 10 96 Source: Drčze and Sen (1995, 1996); Public spending numbers from annual publications of the Reserve Bank of India Bulletin on state finances in India; India numbers on public spending only for 15 major states If spending on health and education is less easy to target than other types of spending, as seems to be the case, then the allocation differences across the two states reflect stronger preferences in UP for targeted spending. The quality of public services is indisputably harder to target and here the evidence is clear: the quality and effectiveness of public spending are higher in Kerala. Drčze and Sen review extensive evidence that public schools and health clinics in Kerala have high teacher and physician attendance, are well supplied, adequately maintained, and widely utilized; similar facilities in UP simply lie abandoned. Why? The formal political and legal institutions in the two states are the same and both states lie within a single sovereign nation. Both have had regular elections to their legislative assemblies since the 1950s , monitored by the Election Commission of India and in accordance with constitutional procedures. Hence the answer cannot lie in differences between political institutions per se, but in the dynamics of political competition. Specifically, Kerala's voters are better informed and less polarized on non-policy dimensions than voters 18 in UP, and its political parties compete on credible platforms for delivering broad social services, as compared to clientelist platforms in UP. Kerala entered the era of democratic elections with a substantially literate, informed, and politically active electorate, as reflected in high turnout rates averaging over 75 percent even in the first few state elections. In contrast, the state of UP entered the new era of democracy with a population of largely illiterate voters with limited access to widely published information, and active institutions of social discrimination which effectively prevented the so-called "backward castes" from accessing public services, or perhaps from participating in political processes. Turnout rates in UP in the first few elections averaged 55 percent. Although caste issues also existed in Kerala, poor and socially disadvantaged voters were more politically mobilized (Nossiter, 1982; Chander, 1986). These affected political party competition and the likelihood that credible political competitors would emerge to contest elections. Following independence from British colonial rule, the Congress Party was the dominant party in India; it had the widest national presence, the most solid reputation, related to its association with the fight for independence, and the longest presence as a political party. Congress espoused a common socialist platform throughout India, focusing on state-led development and redistribution to the poor. However, at the state level there was substantial variation in the perception of the Congress Party by voters, the extent and nature of competition confronting the Congress Party, and the types of credible promises to voters that the party could or had to make to be successful. In many states, such as in UP, Congress did not confront vigorous competition from credible and well-organized parties. In others, credible challenges to the Congress were mounted by regional parties and newly emerging communist parties. In Kerala, the communist parties were particularly active and invested substantially in mobilizing the poorest voters, and by repeatedly interacting with these voters, developed the ability to make credible promise to them (Chander, 1986; Nossiter, 1982). Competition in Kerala was therefore between two credible political parties, the Congress and communist parties, both able to make promises to broad segments of society. This diminished the importance of patron-client relationships, to an extent that was unparalleled in any other state. In UP, on the other hand, massive voter illiteracy undermined the ability of any party to make credible promises to voters, since reputation with voters is difficult to establish if voters are poorly informed. Moreover, in conditions of visible and often coercive social discrimination, organization of the victims of discrimation was in any case highly costly. Two anecdotes highlight the difference in the costs of organizing the poor in the two states. Mencher (1980) describes how in Kerala, if a primary health center were to be unmanned for a few days, there would be massive demonstrations at the nearest district office, with people demanding redress. Drčze and Gazdar (1996) in contrast recount how a village school in UP can be non-functional for as long as ten years due to teacher absenteeism and shirking, without any collective protest being organized. On the one hand, then, the Congress Party could not make credible promises related to its socialist platform of redistribution to the poor in UP. On the other hand, given the absence of challengers, there was no political payoff to the Congress Party from making such promises. In fact, the Congress party won on average more than 70 percent of the seats in the state assembly (numbers based on data provided in Butler, Lahiri, and Roy, 19 1995). In Kerala, the payoff to broad public good promises was high, both because competing political parties could make them credibly, and because neither party could offer the same level of benefits to broadly mobilized constituents through expensive, targeted programs that they could offer through broad-based, high quality service provision. In UP, there was little political payoff to competing on the basis of broad public goods or redistributive programs; narrow, targeted goods and services, despite their limited impact on welfare, were politically superior. The comparisons of both budget allocations and the effectiveness of spending support this conclusion. As Figures 2a and 2b show, Kerala's public resources were overwhelmingly allocated to education and health services in the early decades of democracy, constituting 45 percent of total expenditures, with correspondingly fewer resources (less than 30 percent) spent on state administration. In UP, during this same period, expenditures were concentrated in state administration, comprising almost 50 percent of the total, while health and education received less than 25 percent of public resources.. The circumstances in Kerala distinguished it as well from most Indian states, since on average the major Indian states spent approximately 30 percent of total spending on health and education and over 40 percent on state administration. Moreover, expenditures in Kerala were effective. The first two decades of elected state government saw dramatic improvements in human development indicators. Infant mortality declined about 43 percent, for instance, between 1956 and 1966 (Krishnan, 1991). This decline has largely been attributed to high female literacy (which was also expanding through state programs in education) but was clearly facilitated by access to public primary health care, and state-sponsored programs of infant and child immunization (Zachariah, 1992). Early political success in delivering high quality public health and education bolstered the credibility of political promises in this area and allowed electoral competition on the platform of social service provision to be sustained. No such progress or evolution was evident in UP, where absenteeism and other indications of widespread shortfalls in service delivery suggest that education and health spending served as political patronage in the form of targeted jobs for teachers and health workers. The pattern established early on in UP has proven difficult to change. UP is not immune to the fact that it is only one state in a country with well-developed democratic institutions; these have slowly penetrated the social landscape in the state over the past fifty years, especially through the political mobilization of traditionally repressed lower castes. However, although new parties could, in principle, have attempted to establish state-wide policy reputations, this is costly in social environments such as that in UP, where incumbents have many tools at their disposal to block such entry. Ethnic or caste links to political candidates are the only connections that voters in UP believe will give them access to state resources. As a consequence, new political parties have organized along caste lines and compete on explicitly clientelist platforms ­ on the basis of narrow, targeted promises, the only promises that are politically useful in such an environment. Among the three main parties competing in the state today, the Bharatiya Janata Party (BJP) appeals to upper caste Hindus, the Bahujan Samaj Party (BSP) to so-called backward castes, scheduled castes and tribes, and minority religion groups (hence the complementary set to upper caste Hindus), and the Samajwadi Party (SP) to similarly marginalized groups along the lines of religion and caste. Clientelist politics in Uttar Pradesh is well illustrated by the BSP platform, currently the dominant political party in the state legislature. The party simply does not print any 20 election manifesto to explain its platform. It does, however, publicize the ethnic profile of its candidate list to demonstrate commitment to its single point program of proportional representation for every ethnic group in the bureaucratic institutions of the state (Chandra, 1999). This example of the contrast between Kerala and UP demonstrates that the same formal institutions of democracy can sustain very different forms of electoral competition with substantial impact on the quality of public services supported by the state. Policy interventions and electoral institutions that change the information processing capacity of voters, impact ideological fragmentation, and the capacity of politicians to make credible commitments, can potentially make a significant difference for the process of political competition and hence for the provision of broad social services. Further research is needed to evaluate the impact of ongoing political and institutional reforms in the developing world, and to draw lessons from the existing variation in institutions across countries. Where poor voters are already active in political processes the real issue is that of bolstering the credibility of political candidates to provide broad social services, with a corresponding reduction in existing political pressures to pursue clientelist policies. Part 2: Political institutions and social service delivery The prior sections demonstrate that credibility, information and social polarization have a significant effect on policy and on the incentives of politicians to provide high quality public goods. Political and electoral institutions also influence these incentives, both in and of themselves and in interaction with social institutions. In the benchmark case described earlier, institutions play a small role because voters are identical and elect a single official who determines policy and then implements it. In reality, multiple policy makers make policy on behalf of heterogeneous voters. The electoral and political institutions through which these policy makers are chosen and within which they establish and implement policy nearly always advance some voters' preferences at the expense of others'. The discussion below reviews what we know about several kinds of institutional arrangements: the electoral institutions that shape how politicians respond to citizens; the political institutions that assign authority to different decision makers within each level of government; and political decentralization, which removes authority from national to local level political decision makers. Electoral institutions Electoral institutions have significant effects on social service delivery. To see this, one need merely consider the quality of social services available to citizens whom electoral rules exclude from the franchise. Apartheid South Africa and the southern United States when it operated under infamous Jim Crow laws presented two such situations, and ample evidence shows that non-voters received lower quality social services.10 Engerman and Sokoloff (2002) have documented a more general relationship between initial inequality in the distribution of factor endowments in Latin America, restrictions on the franchise, and low access to education. However, even in democracies with fully-extended franchises, the specifics of electoral institutions generate large differences in policy. This section considers several: the 10 On the access of Southern blacks to education under Jim Crow laws, see Margo (1990). 21 number of voters represented by legislators and variations within country in this number; the number of legislators from a single electoral district (district magnitude); and the type of electoral rule (proportional representation or plurality-based) used. There is substantial variation across countries with respect to these institutions. For example, in 2000, the governments of 98 countries had been elected in competitive elections.11 Of these, about half (49) utilized proportional representation/party list electoral systems and half (47) employed plurality systems. The choices of developing countries deviate from this even distribution, however. Among 32 countries with PPP-adjusted per capita incomes of less than $5,000 (the median income of democracies), only one-third used proportional representation; the other two-thirds (22) used plurality systems. One important electoral institution is simply the extent of disparities within a country in the number of voters represented by legislators. If some legislators represent a large number and some a small number of voters, the latter voters are (all else equal) likely to exercise greater influence on policy. These variations are, in fact, quite large, and seem to account for certain significant policy outcomes in countries where they are largest. In the Dominican Republic and the United States, sparsely populated regions are given the same representation in one of the legislative chambers as more populated regions. The state of California, with a population of approximately 33 million people, has the same number of senators (two) as each of the 24 smallest states, which together have approximately 36 million residents. The Distrito Nacional of the Dominican Republic has 28 percent of the electorate and one senator; the 16 smallest provinces each have one senator, as well, but together they have only 23 percent of the electorate. In these two countries, the senators from the largest electoral districts represent twenty times (or more) as many voters as senators from the smallest electoral districts. There is a similar pattern, but perhaps even more extreme, in India. The largest constituency of the lower house, the Lok Sabha, has 25,000,000 voters, while the smallest has only 50,000. In Canada, on the other hand, most electoral districts have approximately 90 ­ 100,000 voters; the smallest has 27,000 and the largest 115,000, a little more than four times larger.12 The over-representation of voters from sparsely populated areas has clear implications for the effects of government policy. For example, since rural areas are more agricultural, over-representation of rural areas leads to more generous government policies towards agriculture than would otherwise be the case. This is evident in all of the countries mentioned above.13 However, over-representation of rural areas skews social service provision, as well. Lee (1997) shows that in the United States, small states receive a disproportionately large share of almost all non-discretionary redistributive transfers, independent of need ­ despite the fact that the small state bias is only strong in the Senate. Per capita federal 11 Elections in which there were multiple competing candidates or parties, more than one party contesting, and no candidate or party winning more than 75 percent of the vote. Information from the Database of Political Institutions (see Beck, et al., 2001). 12 Information from Keefer (2002 XX), Elections Canada On-Line www.elections.ca, and the Electoral Commission of India (http://www.eci.gov.in/). 13 Itis important to emphasize that, depending on the information and credibility issues discussed earlier, the influence of rural areas in policy making need not mean that the average rural voter benefits, only that the politically influential rural interests benefit (who may or may not be the median rural voter). 22 outlays for nondiscretionary, distributive spending (such as community development, transportation subsidies, education subsidies and employment and labor training programs) were as high as $209 for small states (Wyoming), compared to$139 for large states (California). US electoral institutions therefore effectively raise the cost of targeting the neediest recipients of such transfers, since such recipients can only be reached by providing larger payoffs than would otherwise be the case to over-represented voters. Cox (1997) describes the importance of these and other dimensions along which electoral institutions vary. voters can cast many votes or a few; district magnitudes (the number of legislators elected by an electoral district) can be as large as the size of the legislature (as in the Netherlands or Israel, where the entire country is a single electoral district) or only one, as in the United States House of Representatives or the British Parliament, where each electoral district or constituency returns only one legislator; and voters must select either party lists or individual candidates. Permutations along each of these dimensions affects political incentives to appeal to any particular group of voters. To the extent that the beneficiaries of quality social services are not in the groups that politicians have an incentive to attract, social services are correspondingly under-provided. There is little in the literature linking the number of votes that voters cast to policy outcomes so the focus of the discussion here is on district magnitudes and proportional representation versus first past the post voting systems. Cox and McCubbins (2001) point out that electoral rules first determine whether voters cast their ballot on the basis of party identity or candidate identity and, second, they determine whether parties have strong or weak incentives to coalesce or to splinter. Systems that encourage voters to choose candidates encourage candidates to push for constituency- specific benefits that they alone can take credit for ­ similar to the earlier arguments regarding clientelism. Systems that encourage parties to splinter similarly encourage parties, once in office, to under-provide broad public goods at the expense of government programs that benefit their narrow constituency base. In many cases, though, the same institutions do both, generating offsetting effects. The United States and Britain, for example, have low district magnitudes (one), encouraging candidate specific voting behavior, but also driving down the number of parties. Persson and Tabellini (2000, chapter 8) make strong policy predictions based on two particular institutional combinations: proportional representation systems with a single electoral district, versus an electoral system with single member districts.14 They argue that the former system offers greater incentives to elected officials to provide broad-based public goods since it forces political competitors to care about voters who are ideologically indifferent between the competing parties and about some voters who are more ideologically committed to one party or the other. The second, winner-take-all system gives competing parties an incentive to focus all of their resources and political promises on swing districts, to the exclusion of districts that are more ideologically committed to one or the other party. This focus reduces political incentives to provide public goods to all voters including, implicitly, high quality social services.15 14 As dis trict magnitudes fall, electoral systems become increasingly "winner-take-all" or "majoritarian", regardless of whether they use proportional representation or plurality voting rules. In the limit, when district magnitudes are one, there is no difference between the two voting rules. 15 The analysis of Persson and Tabellini (2000, chapter 8) generates this result assuming credible 23 What effect do these differences have on policy? With respect to the impact of electoral rules, Persson and Tabellini (1999) draw a distinction between systems with district magnitude equal to one and all other systems and they define public goods as expenditures on transportation, education, and public order and safety (and, in an alternative measure, health spending). With these assumptions, they find some econometric evidence consistent with their argument that low district magnitudes drive political competitors to focus their promises on a smaller number of voters, reducing political incentives to provide public goods. They find less evidence that there is a bias in presidential versus parliamentary systems in the allocation of resources across budget categories, but there is very strong evidence that overall spending in presidential systems, including presumably on broad public goods and social services, is much lower than in parliamentary systems. Political institutions In all democratic countries, policy is made by multiple elected politicians. Decision making authority is rarely distributed evenly among them across all issues. On the contrary, the distribution of decision making authority varies across countries along three important dimensions: which politicians can set the agenda; which can veto proposed changes in law or regulation; and which can force other politicians to leave office or to seek re-election. All have an impact on policy. For example, agenda setting authority is the power to make proposals for consideration by other policy makers. The stronger are the veto and agenda- setting powers of political decision makers, the more that policy will reflect their interests and those of their constituents to the exclusion of others in the society. In Chile and many other Latin American countries, only the president can propose the national budget. In the United States, on the other hand, only the House of Representatives can originate a spending bill. In most legislatures, the opposition of a majority of the representatives is sufficient to veto a bill; in the US Senate, on the other hand, procedural rules have the effect of giving any 40 percent of the Senators the opportunity to block legislation. Finally, some countries' executives can call new elections for the legislature (e.g., Russia), while in other countries the reverse is true: legislatures can bring down the government without having to go through new elections (e.g., Italy). The greater is the ability of one political actor to force, at minimal risk to herself, another political actor to leave office or to seek re-election, the more that her policy preferences and those of her constituents will be reflected in policy making. Two frequently studied sets of political institutions are parliamentary and presidential systems. These can be thought of as simply two bundles of institutional arrangements governing the assignment of veto and agenda-setting power and the control of the executive and legislature over each other's electoral destinies. Forty-three percent of all democracies in 2000 were presidential, but of the 32 poorest democracies, 69 percent were presidential. District magnitudes are also quite different. Among all democracies, the median district magnitude was almost seven ­ but was only four among poor democracies. However, among democracies with proportional representation electoral systems, median district magnitudes are similar in developing and all democracies (7.4 and 9.2, respectively). politicians and fully informed voters. The result depends in addition on ideological differences between parties and imperfect information on the part of politicians about the ideological preferences of specific voters. 24 Persson and Tabellini (2000) model parliamentary systems as cabinet government in which cabinet members have exclusive, all-or-nothing proposal power over their portfolios, one portfolio relates to spending and the other to taxation, and all cabinet members are entitled to veto the proposals of all other cabinet members. In presidential systems, all proposal power rests with the legislature (the executive can make no amendments, but can only disapprove or approve the final package), and proposal power within the legislature is dispersed, as in the cabinet. However, the committee in charge of tax proposals cannot veto the proposals of the spending (appropriations) committee, and vice versa. Instead, proposals are rejected only if a majority of the legislature vote them down. Finally, Persson and Tabellini assume that both systems operate under the umbrella of a majoritarian electoral system. Although apparently obscure, these different institutional arrangements have striking implications. Because the tax committee in the presidential system cannot veto the spending committee's proposed allocation of spending, it knows that it will have to accept a lower spending allocation to its own constituents than it would otherwise be able to extract. As a consequence, the tax committee sets taxes very low. This drives rent-seeking and targeted transfers to specific voters down to zero, but also severely reduces public good provision. In parliamentary systems, though, the institutional set-up guarantees the tax minister that he will get a large allocation, so he proposes a high tax rate. Public good provision, targeted transfers and rent-seeking are all high in parliamentary systems. There is mixed and not entirely robust evidence in support of these predictions. Persson and Tabellini (1999) find weak evidence of a bias in presidential versus parliamentary systems in the allocation of resources across budget categories. However, there is strong and robust evidence that presidential systems spend much less than parliamentary systems. The weak empirical results are nevertheless instructive because they are a reminder to revisit the institutional assumptions underlying the comparison of presidential and parliamentary regimes. The power of individual cabinet members in parliamentary government to veto the proposals of other cabinet members depends, for example, on whether a vote of confidence procedure is in place. If it is, as Diermeier and Fedderson (1998) argue, then a vote against the bill means that the government falls. If the government will fall if a cabinet member opposes a bill, then one might expect substantial efforts to ensure every cabinet member is happy with the bill in question. However, in every country with a vote of confidence procedure, it is the prime minister who must propose that a vote on a bill be a confidence vote (Huber 1996). In practical terms, therefore, the vote of confidence procedure increases the bargaining leverage of the prime minister over recalcitrant cabinet members. This has the effect of centralizing proposal and veto power in the hands of the prime minister. There are at least two reasons why parliamentary systems would not necessarily generate policies predicted by Persson and Tabellini. First, parliamentary democracies may not have a vote of confidence or, more likely, the vote of confidence may not be credible. Huber looks at 18 OECD parliamentary democracies with votes of confidence and finds that in only six is it written into the constitution. In the other cases, the vote of confidence is based on convention or standing orders of parliament, with no or only modest legal obstacles to being ignored. In other countries, there may be no vote of confidence at all. Absent the vote of confidence, though, we do not expect to see legislative cohesion in 25 parliamentary systems; in particular, we expect veto power to reside with the whole legislature, just as it does in the presidential system. Second, if some coalition partners expect to be able to join successor governments with ease, then the prime minister will never declare a vote on a particular bill opposed by these coalition partners to be a vote of confidence. To do so would be to guarantee that the government would fall, since the recalcitrant coalition partners would lose nothing by voting against the bill. This is not an uncommon occurrence. In India, the AIADMK party was in the coalition headed by the Congress Party from 1992 to 1996, but when this government fell and the Congress Party moved into opposition, in 1999 and 2000, the AIADMK party entered the coalition led by the BJP. Although there are few poor countries that are parliamentary democracies (eight in 2000), similar results can be found for these, as well (e.g., Papua New Guinea) and for Western European democracies (e.g., in the Netherlands), where some parties have regularly managed to be part of successive governments. In such countries, again, we expect policy outcomes closer to those of the presidential system. Presidential systems also exhibit institutional variation that leads to departures from the predictions of the stylized Persson and Tabellini model. The most important is the role of the president in budget preparation. In the United States, the model for Persson and Tabellini's presidential-congressional system, tax and spending bills originate in different Congressional committees and the president can only exercise an all-or-nothing veto over legislation. This is not the case in many other presidential democracies. In nearly all Latin American presidential democracies, only the president can propose the budget. In many of these, as Keefer (2002a) finds, the legislature confronts severe restrictions on the amendments it can make to the proposal (e.g., Perú and Colombia) and the President's proposal takes effect if the legislature fails to agree on a new budget (e.g., Bolivia, Colombia, Chile, Ecuador and Perú). If the president also enjoys similar authority over tax legislation ­ which is not clear from the available evidence ­ then one would expect outcomes similar to those projected for parliamentary systems. Finally, presidential democracies differ in the control presidents exercise over the legislature. For example, in Russia the president can call for new legislative elections without himself having to run again for office, reducing the veto power that the legislature can exercise over presidential proposals. In the end, if it is the case that parliamentary systems generally exhibit a vote of confidence and there are generally high costs to government coalition members of bringing down the government, and if it is the case that budget and spending are split between two different authorities that lack mutual veto power in presidential systems, then we expect parliamentary systems to provide significantly more public goods (and to exhibit significantly higher transfers and corruption) than presidential systems. Social service provision, in particular, should be higher in parliamentary systems. To the extent that the vote of confidence feature is weak or missing in the parliamentary systems, however, public good provision in parliamentary systems should fall; to the extent that presidents have strong agenda control over both budget and tax legislation, on the other hand, public good provision should be higher. Much more work is needed to document the impact of political and electoral institutions on economic outcomes generally and social service delivery specifically. The discussion above provides specific clues about where such work is most needed: in disaggregating and more finely identifying the institutional arrangements that shape political incentives and drive political decision making. In addition, however, future work is needed 26 to assess the interaction with formal institutions of the conditions under which political competition occurs (the extent to which voters are informed and ideologically polarized, or to which politicians can make credible promises to voters). Political Decentralization Another important institutional change, particularly emphasized because of its potential impact on the poor, is the decentralization of political authority to local, especially village governments. Proponents of decentralizing responsibility for many social services to the local level argue that by bringing government physically closer to the people, and narrowing the scope of public activities for each tier of government, voters will find it easier to hold politicians accountable for performance. Roughly speaking, decentralization improves outcomes to the extent that physical proximity increases voter information about performance, and to the extent that narrowing the scope of responsibilities of each tier of government decision makers reduces their ability to shirk on some responsibilities by performing better on others. More generally, and consistent with the earlier discussion, local governments will provide better quality public goods than regional or national governments if voters are better informed and likely to use information about local public goods in their voting decisions for electing local governments, if there is greater social homogeneity and coordination of preferences for local public goods, and if political promises are more credible at local levels.16 In addition, if voters feel strongly about multiple issues, some of which are then decentralized to local governments, performance should also improve.17 Not all of these factors need to operate in favor of decentralization. Voters may be better informed about the quality and availability of local public goods because of greater physical proximity, or more focused on using this information in voting decisions because of the narrower range of responsibilities for which to hold their representatives responsible. However, information with regard to local policies may be of poorer quality if national newspapers, covering only national issues, are the main source of information for voters. Furthermore, if local expenditures are financed entirely out of grants from higher tiers of governments and not out of local tax bases (as is the case in most developing countries), then local voters may have little or no information regarding the resource envelope available to their local government and what those resources are intended to provide. 16 The political institutional reforms discussed here involve decentralization to the level of well-defined communities--close-knit villages and municipalities--that share local public goods for which locally elected representatives are made responsible. This section does not deal with federal forms of government where some authority and responsibility is decentralized to relatively large regional governments, because bulk of the arguments for decentralization are based on proximity of the elected government to the electorate, which should be of a smaller dimension than what would exist in a regional, although sub- national government. 17 The nature of voter preferences is a new theme, not previously discussed. The idea is simply the following. If voters care deeply about both education and national defense, national government decision makers can more easily remain in office by doing well on the second and under-peforming on the first. Decentralizing responsibility for education to a lower level government allows voters to hold one set of officials strictly accountable for education and the upper tier strictly accountable for defense. If voters care only about education, however, then local and national decision makers would have similar incentives to perform well on education, and decentralization would have little effect. 27 Social homogeneity may also decline rather than increase at more local levels. Social polarization between any two local groups may be more intense due to age-old differences across settled communities, and weaker at the national level, perhaps through national campaigns of nation-building. Local politics may therefore be more likely to revolve around identity issues and hence not geared towards providing strong incentives for political agents to deliver public goods. Political agents at appropriately decentralized levels may have greater credibility to voters at large because of their proximity or reputation developed through community interactions over an extended period of time. However, these same features may allow clientelist promises to be easier to make and fulfill at more local levels due to closer social relations between the elected representatives and their clients, at the expense of broad public goods. No systematic evaluations of decentralization take into account these features of political competition. These make it difficult to use decentralization assessments from one setting to draw conclusions about likely effects in other settings. Nevertheless, emerging evidence from decentralization reforms in three diverse parts of the developing world-- India, Bolivia, and Nigeria ­ shed important insights in their own right. Decentralization reforms in India were instituted by a constitutional amendment in 1993. They are intended to transfer responsibility from state governments to newly created locally elected village councils. The amendment also calls for participatory citizen decision- making bodies at the village level (Gram Sabhas) and political reservation for women and traditionally disadvantaged groups (scheduled castes and tribes) on the village councils. Indian states were, however, afforded considerable discretion in the design and timing of these decentralization reforms leading to considerable variation across states and villages in the extent to which these reforms have been implemented. Foster and Rosenzweig (2001) conducted one of the first evaluations of the impact of decentralization on government performance. They use a panel of 250 villages in rural India with information on economic characteristics, public good stocks, and local political structures observed before and after significant political decentralization to estimate the impact of local democratization on village-level provision of public services. They focus on 3 categories of public goods which together account for 73 percent of the activities of village governments in their sample in India--roads, irrigation, and schools. They find that villages with democratically elected governments are more likely to provide more of all three public goods, but the largest effect is for irrigation, as calculated at the sample average. However, in villages with a very high proportion of landless (much above the sample average) public investment shifts from irrigation to road construction (rather than education, which is unaffected by proportion landless). Their findings suggest that decentralization has an ambiguous affect on poverty. Roads built by village governments primarily benefit the poor, but largely by raising their (short-term) wages, as local road construction and improvement initiatives in India serve as employment programs for the landless poor. Irrigation facilities disproportionately benefit the richer landowners, however. Moreover, education, which one expects to have the most profound effect on poverty over the medium and long-term, seems least affected by decentralization. In similar vein, in a study of villages in the states of West Bengal and Rajasthan, Chattopadhyay and Duflo (2003) find that villages with women leaders, elected as a 28 consequence of mandated political reservations for women in village governing bodies, are more likely to invest in public goods that are revealed preferred by women during interviews--water, fuel and roads (as it provides employment) and less likely to invest in education. They also find greater participation by women in citizen decision-making bodies in West Bengal if the village government is headed by a woman. Their findings do not necessarily reflect the success of decentralization, per se, but rather of political reservations at any level of government in environments where there are substantial social pressures operating against the political participation of particular groups. Taken together, the work on political reservations and village decentralization in India suggests that traditionally disadvantaged groups that receive new democratic privileges tend to exert pressure to shift public resources out of education and into other targeted public goods that provide immediate benefit specifically to their group. While this suggests that political decentralization coupled with political reservations indeed succeeds in giving greater voice to these disadvantaged groups, it is also indicative of problems in the public provision of education services.18 Decentralization in India may also have improved the efficiency of local public services. Quite apart from the impact of decentralization reforms on intersectoral allocation of public resources, Crook and Manor (1994) offer qualitative evidence that decentralization of the management of primary health and education to village level governments in the state of Karnataka improved service delivery outcomes by considerably reducing absenteeism and shirking amongst teachers and health workers. In Bolivia, dramatic political reforms were undertaken in 1994 with the creation of democratically elected local governments and participatory decision-making bodies where none existed previously. While prior to these reforms municipal governance bodies had existed in the major cities of Bolivia, the 1994 Law of Popular Participation (Ley de Participación Popular) created new local governments in rural areas. It required untied grants, amounting to 20 percent of national tax revenues, to be distributed among all local governments on a per capita basis to fulfill their responsibilities in the provision of local public goods in health, education, and infrastructure. In addition, participatory bodies with representatives from community organizations, known as Oversight Committees (Comites de Vigilancia), were established with substantial legal powers to oversee the allocation and spending of these municipal grants by the local governments. In a detailed study of the impact of these political reforms Faguet (2001) finds that they were responsible for two major changes in the pattern of allocation of public resources in Bolivia. First, there was a sharp fall in the geographic concentration of public investment as investments became more equally spread across regions. Second, a significant shift in the allocation of public investment took place, away from heavy infrastructure and towards the social sectors (e.g., school buildings). Furthermore, variation across municipalities shows that municipal investment patterns are responsive to local needs, with greater investment in 18 The shift away from education may be just a short-term effect of decentralization reforms. The immediate concern of new women political leaders, for example, may be to provide those essential public goods that disproportionately benefit women and have been historically underprovided. Demand for public resources in education may increase with time, as women are better situated to take advantage of improved opportunities. A full analysis of the dynamic impact of political decentralization in India, as the institutions stabilize over time, remains to be undertaken. 29 the social sectors precisely in those municipalities where education and health indicators were of poorest quality at the onset of decentralization, and greater investment in urban infrastructure where there was a larger concentration of private firms. In Nigeria, with the fall of military dictatorship in 1999, democratic elections were simultaneously held for local governments in the country, as at the federal and state levels. Local governments in Nigeria are responsible for basic services in health and education and for local infrastructure, and have constitutional rights to federally collected revenues, regularly receiving general purpose fiscal transfers (about 25 percent of total shareable revenues of the federation), for the discharge of their expenditure responsibilities. Yet, a scrutiny of some local government budgets shows that resources are largely devoted to capital infrastructure projects, with little allocation for operations and maintenance of schools and clinics (IMF, 2001). Field-work in local government areas in one state revealed pervasive non-payment of salaries of health personnel reporting to the local government authorities, with consequent staff absenteeism and poor quality of service delivery (Adeniyi and Oladepo, 2003). Non-payment of teachers' salaries by local governments had been a persistent problem through the 1980s and 1990s leading to organized agitation by national teacher unions, and resulting in direct payment of teachers' salaries by the federal and state government through a deduction at source from the local governments' share in federal revenues. The experience with local government in Nigeria is therefore a striking contrast to the Indian and Bolivian experience described above. Yet, the comparison may not be reasonable given that no study in Nigeria has actually addressed the issue of the impact of decentralization by comparing outcomes in more and less decentralized regions, or over time as more or less resources are managed by local governments. On-going studies in Nigeria are attempting this. One lesson emerging from this work is that the history of centralization of resources and responsibilities in Nigeria has led to confusion and lack of information at local levels with regard to what local agencies are actually responsible for and what fiscal transfers they receive. Local governments have so far tried to feign fiscal powerlessness, making it easier for them to divert public resources away from the services for which they were intended (IMF, 2001). It remains to be seen whether this process will continue after local elections become a regular feature of the Nigerian political landscape. Conclusion: Reforms to soften distortions in political and electoral markets As the Millenium Development Goals make clear, development policy has turned ever more emphatically to a focus on improving the condition of the poor. Essential to that effort are the quality and quantity of social services provided to the poor by governments. At the same time, and in parallel, there is an increasing awareness that "institutions matter" for development. Moreover, recent, extensive and rigorous work by social scientists has revealed a great deal about the impact of electoral competition and a broad array of political and electoral institutions on policy outcomes. In this paper, we attempt to join these two parallel lines of argument. What can we derive from the political economy literature about the impact of electoral contestation and political decision making on social service outcomes to the poor? Theory and evidence suggests that the impact is large, and largest for those social services most important to the poor, education and health. Information gaps, social polarization and the absence of credible political competitors lead to the underprovision of 30 government services to the least informed, to the most polarized and to the vast majority who do not receive benefits from a powerful patron. Even when services are provided, they less often take the form of high quality public services available to all. Electoral and political institutions exacerbate these distortions when they encourage politicians to provide targeted services to narrow groups of constituents or when they dilute the electoral strength of the poorest voters. Decentralization to the village level, though often wresting power from national and state level officials with little regard for the poor ­ or, in fact, for any broad public interest ­ can improve the distribution of resources and give the poor greater access to the benefits of government spending. However, it has proven to be no panacea in shifting the pattern of public spending away from targeted spending (jobs, construction) to broad-based, high quality public goods that potentially offer greater benefits to more of the poor at lower cost (high quality education, for example). There are numerous steps that can be taken to begin, at least, to mitigate the effects of particularly distorted political and electoral markets. Voter information can be improved through interventions that provide independent validation about the quality of public goods and the scope of accomplishment and failure of individual political decision makers. The press provides one source of such validation, but so also do external agencies, civic groups, and even competing parties. Survey instruments, such as citizen report cards that consolidate public feedback on the state of government services, are potentially powerful vehicles for mobilizing voters around the issue of quality of public services. There are examples of particular experiences from around the globe of how such "information campaigns" have succeeded, but we have no evidence for whether instruments of this kind could significantly and systematically alter the nature of political competition, without which there would be no lasting impact on policy outcomes. Further research would be valuable on how information provision mechanisms can be institutionalized to enable voters to provide stronger incentives for politicians to improve performance.19 We need to understand as well the feasibility of such reforms. In a competitive political environment, reforms to enhance information have the potential to be accepted, since it can be politically costly to resist them. However, in a non-competitive political environment ­ one where information problems have generated a large incumbent advantage, for example ­ it would be excessively optimistic to expect such reforms as enhancements to the freedom of the press to be encouraged or to be effective. Credibility, like information, is particularly likely to be problematic in developing countries and also requires special attention. Outside interventions that help politicians both implement and take credit for broad public good improvements can lead to sustainable improvement in public good provision, if they help politicians build a reputation for performance. However, clientelist transfers will always be an available low risk political strategy in these environments and intervention designs need to circumvent this. More innovative research is required, both at the sector level , as in high-impact interventions in health and education services which politicians might find easier to commit to and take 19 Evidence concerning the role of the media naturally gives rise to the question: under what conditions is free (or low cost) media informative? Several authors have examined the conditions of competition in political and economic markets and the resulting incentives for the kind and quantity of information disseminated by the media industry (Stromberg, 2002; Djankov et al, 2002; Mullainathan and Shliefer, 2002). 31 credit for, and at the political economy level, as in unpacking the notion of "credibility" and understanding its process of change. Finally, social polarization can have harmful effects on social service provision, just as on other aspects of civic life. We need a better understanding of the root causes of polarization and the emergence of a correspondence between voting behavior and social group membership. The response to pure taste-based affiliations (voting systematically for the representatives of one particular ethnic group or tribe out of an exogenous and strong preference for anyone from that group or tribe over any other candidate) demands educational responses; if polarization is due to tremendous imperfections in political and economic markets, however, the appropriate response is to assist in reforms that remove those imperfections. 32 References Adeniyi, J. and Oladepo, O. (2003). Survey of Primary Health Care Service Delivery in Nigeria: Field Report. Mimeo. University of Ibadan Alesina, A., R. Baqir and W. Easterly (1999). "Public Goods and Ethnic Divisions". The Quarterly Journal of Economics, 1243-1284 (November) Ballou, Dale (1996). "Do Public Schools Hire the Best Applicants?" The Quarterly Journal of Economics 111:1, 97 ­ 133 (February). Banerjee, A. and Somanathan, R. (2001). "Caste, Community and Collective Action: The Political Economy of Public Good Provision in India." Mimeo. Department of Economics, MIT. Baron, David P. (1994). "Electoral Competition with Informed and Uninformed Voters". American Political Science Review, 88, 33-47 (March) Bates, R. (1981). Markets and states in tropical Africa: The political basis of agricultural policies. University of California Press, Berkeley. 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. Bertrand, M. and S. Mullainathan (2001). "Do CEOs Set Their Own Pay? The Ones Without Principals Do". Quarterly Journal of Economics, 116, Besley, T. and R. Burgess (2003). "The Political Economy of Government Responsiveness: Theory and Evidence from India". Quarterly Journal of Economics, Forthcoming Betancourt, R. and Gleason, S. (2000). "The Allocation of Publicly Provided Goods to Rural Households in India: On Some Consequences of Caste, Religion and Democracy". World Development, 28(12), 2169-2182 Block, S. A. (2002). "Political business cycles, democratization, and economic reform: the case of Africa". Journal of Development Economics, 67, 205-228. Butler, D., A. Lahiri, and P. Roy. (1995). India Decides: Elections 1952-1995. New Delhi. Campbell, A., P. Converse, W. Miller, and D. Stokes. (1966). The American Voter. New York: J. Wiley & Sons. Card, David and Alan B. Krueger (1992). "Does School Quality Matter? Returns to Education and the Characteristics of Public Schools in the United States." Journal of Political Economy 100: 1, 1-40 (February). 33 Carlson, Beverly A. (2000). "Achieving Educational Quality: What Schools Teach Us ­ Learning from Chile's P900 Primary Schools." Serie Desarrollo Productivo No. 64, CEPAL, Santiago, Chile (January). Chander, Jose N. ed. (1986). Dynamics of State Politics in Kerala. Sterling Publishers, New Delhi Chandra, K. (1999). "The Ethnification of the Party System in Uttar Pradesh and its Consequences." In R. Ramashray and P. Wallace (eds.) Indian Politics and the 1998 Elections: Regionalism, Hindutva, and State Politics, Sage Publications, London Chappell, H. and W. Keech (1990). "Citizen Information, Rationality, and the Politics of Macroeconomic Policy". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Chattopadhyay, R. and E. Duflo (2003). "Women as Policy Makers: Evidence from a India-Wide Randomized Policy Experiment". MIT Department of Economics, Working Paper Clague, Christopher, Philip Keefer, Stephen Knack, Mancur Olson (1996). "Property and Contract Rights under Democracy and Dictatorship." The Journal of Economic Growth 1:2, 243-276 (June).[NB1] Cox, Gary (1997). Making Votes Count: Strategic Coordination in the World's Electoral Systems. New York: Cambridge University Press. _____ and Mathew McCubbins (2001). "The Institutional Determinants of Economic Policy Outcomes." In Presidents, Parliaments and Policy, Stephan Haggard and Mathew McCubbins, eds., 21-63. New York: Cambridge. Crook, Richard and James Manor (1994). Enhancing Participation and Institutional Performance: Democratic Decentralization in South Asia and West Africa. Overseas Development Administration, London. Diermeier, Daniel and Timothy Fedderson (1998). "Cohesion in legislatures and the vote of confidence procedure." American Political Science Review 92: 611-21. Djankov, Simeon, Caralee McLiesh, Tatiana Nenova, and Andrei Shliefer (2002). "Who Owns the Media?" Journal of Law and Economics, Forthcoming Downs, Anthony (1957). An Economic Theory of Democracy. New York: Harper and Row. Drčze, Jean and Amartya Sen (1995). India: Economic Development and Social Opportunity, Oxford University Press 34 Drčze, Jean and Amartya Sen (1996). Indian Development: Selected Regional Perspectives, Oxford University Press Drčze, Jean and Haris Gazdar (1996). "Uttar Pradesh: The Burden of Inertia". In Drčze, J. and A. Sen (eds.) Indian Development: Selected Regional Perspectives, Oxford University Press. Easterly, W. and R. Levine (1997). "Africa's Growth Tragedy: Policies and Ethnic Divisions". The Quarterly Journal of Economics, 112(4), 1203-1250 Ehrenberg, Ronald G., Randy A. Ehrenberg, Daniel I. Rees, Eric L. Ehrenberg (1991). "School District Leave Policies, Teacher Absenteeism, and Student Achievement." Journal of Human Resources 26:1, 72-105 (Winter). Engerman, Stanley L. and Kenneth Sokoloff (2002). "Factor Endowments, Inequality, and Paths of Development Among New World Economies." NBER Working Paper 9259 (October). Faguet, J. (2001). "Does Decentralization Increase Responsiveness to Local Needs? Evidence from Bolivia". World Bank Policy Research Working Paper No. 2516, The World Bank, Washington DC Ferejohn, John (1986). "Incumbent performance and electoral control." Public Choice 50: 5-26. _____ (1990). "Information and the Electoral Process". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Ferejohn, John and James Kuklinski (1990). Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Feyzioglu, Tarhan, Vinaya Swaroop and Min Zhu (1998). "A Panel Data Analysis of the Fungibility of Foreign Aid." World Bank Economic Review 12:1 (January), 29- 58. Fiorina, M. (1990). "Information and Rationality in Elections". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago _____ and K. Shepsle (1990). "A Positive Theory of Negative Voting". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Foster, A. and M. Rosenzweig (2001). "Democratization, Decentralization, and the Distribution of Local Public Goods in a Poor Rural Economy." Working Paper, Brown University, http://adfdell.pstc.brown.edu/papers/democ.pdf 35 Glaeser, E. and A. Shliefer (2002). "The Curley Effect". NBER Working Paper No. 8942 Glewwe, Paul and Hanan Jacoby (1994). "Student Achievement and Schooling Choice in Low-Income Countries: Evidence from Ghana." The Journal of Human Resources 29:3, 843-864. Grossman, G., Helpman, E., (1996). "Electoral competition and special interest politics". Review of Economic Studies 63, 265-286. Hanushek, Eric A. (1996). "Measuring Investment in Education." Journal of Economic Perspectives 10:1, 9-30 (Fall). Harber, Clive (1989). Politics in African Education. London: Macmillan. Huber, John (1996). "The Vote of Confidence in Parliamentary Democracies." American Political Science Review 90: 2, 269-82. Hungerford, Thomas and Gary Solon (1987). "Sheepskin Effects in the Returns to Education." Review of Economics and Statistics 69:1, 175-77. International Monetary Fund (IMF). (2001). Nigeria: Options for Reforming Intergovernmental Fiscal Relations, Washington, D.C. Iyengar, S. (1990). "Shortcuts to Political Knowledge: The Role of Selective Attention and Accessibility". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Keefer, Philip (2002a). "The political economy of public spending decisions in the Dominican Republic: credibility, clientelism and political institutions." Report prepared for the Public Expenditure and Institutional Review, The Dominican Republic. The World Bank. _____ (2002b). "Clientelism, Credibility and Democracy." ____ and Stephen Knack (2002). "Boondoggles and Expropriation: Rent-seeking and Policy Distortion when Property Rights are Insecure." Policy Research Working Paper 2910, The World Bank, (October 11). Khemani, Stuti (2003) "Political Cycles in a Developing Economy: Effect of Elections in the Indian States". Journal of Development Economics, forthcoming. Knack, Stephen and Philip Keefer (1997). "Does Inequality Harm Growth Only in Democracies? A Replication and Extension." American Journal of Political Science 41:1 (January). Krishnan, T. N. (1991) "Kerala's Health Transition: Facts and Factors." Center for Population and Development Studies, Harvard University. Layard, Richard and George Psacharopoulos (1974). "The Screening Hypothesis and the 36 Returns to Education." Journal of Political Economy 82:5, 985 ­ 998. Lee, Frances (1997). "Representation and Public Policy: The Consequences of Senate Apportionment for the Geographic Distribution of Federal Funds." Journal of Politics 60:1, 34 ­ 62 (February). Mani, A. and S. Mukand (2002). "Democracy, Visibility and Public Good Provision". Mimeo. Williams College and Tufts University. Margo, Robert A. (1990). Race and Schooling in the South, 1880 ­ 1950. Chicago: University of Chicago Press. Mayhew, David (1974). Congress: The Electoral Connection. New Haven: Yale University Press. Meltzer, A. and S. Richard (1981). "A Rational Theory of the Size of Government". Journal of Political Economy, 89:5, 914-927. Michaelowa, Katharina (2001). "Primary Education Quality in Francophone Sub- Saharan Africa: Determinants of Learning Achievement and Efficiency Considerations." World Development 29:10, 1699-1716. Miguel, E. (2001). "Ethnic Diversity and School Funding in Kenya". Working Paper C01-119, Center for International and Development Economics Research, University of California, Berkeley. Mullainathan, S. and A. Shliefer (2002). "Media Bias". Harvard Institute Research Working Paper No. 1981, and MIT Department of Economics Working Paper No. 02-33 Nossiter, T. J. (1982). Communism in Kerala : A study in political adaptation. Oxford University Press, Delhi Ottati, V. and R. Wyer (1990). "The Cognitive Mediators of Political Choice: Towards a Comprehensive Model of Political Information Processing". In John Ferejohn and James Kuklinski (eds.) Information and Democratic Processes. University of Illinois Press, Urbana and Chicago Ozler, B., G. Datt, and M. Ravallion (1996). "A database on poverty and growth in India". Mimeo, Policy Research Department, World Bank, Washington, DC. Pande, Rohini (2003). "Can Mandated Political Representation Increase Policy Influence for Disadvantaged Minorities? Theory and Evidence from India". American Economic Review, Forthcoming Persson, Torsten and Guido Tabellini (1999). "The Size and Scope of Government: Comparative Politics with Rational Politicians. 1998 Marshall Lecture" European Economic Review 43: 699-735. 37 _____ (2000). Political Economics: Explaining Public Policy. Cambridge, MA: The MIT Press. Prasad, K. V. Eswara (1987). Wastage, Stagnation, and Inequality of Opportunity in Rural Primary Education: A Case Study of Andhra Pradesh, Ministry of Human Resource Development, New Delhi Reserve Bank of India Bulletin. Reserve Bank of India, various issues Robinson, James and Thierry Verdier (2002). "The Political Economy of Clientelism." CEPR Working Paper 3205, February. Sanmartín, Marta (2001). "Linearity of the return to education and self-selection." Applied Economics 33 (103 ­ 142). Schuknecht, L. (2000). "Fiscal policy cycles and public expenditure in developing countries". Public Choice, 102, 115-130. Shi, M., Svensson, J. (2000). "Political budget cycles: Do they differ between developed and developing countries?" Mimeo, Development Research Group, World Bank, Washington, DC Stromberg, David (2001) "Radio's Impact on New Deal Spending", mimeo, Princeton University. Stromberg, David (2002) "Mass Media Competition, Political Competition, and Public Policy", Review of Economic Studies, Forthcoming. Tilak, Jandhyala B. G. (1993). "Costs and Financing of Education in India: A Review of Issues, Problems, and Prospects". Center for Development Studies, Trivandrum Troesken, Werner (2002). "The Limits of Jim Crow: Race and the Provision of Water and Sewerage Services in American Cities, 1880-1925." The Journal of Economic History 62: 3, 734-772. Van de Walle, Dominique and Kimberly Nead (1995). Public Spending and the Poor: Theory and Evidence. Published for The World Bank, Johns Hopkins University Press, Baltimore and London Weiner, M. (1991). The Child and the State in India: Child Labor and Education Policy in Comparative Perspective, Princeton University Press, Princeton Weiner, M. and J. O. Field, (1974). Studies in Electoral Politics in the Indian States, Delhi: Manohar Book Service Wolfers, Justin (2002). "Are Voters Rational? Evidence from Gubernatorial Elections". Research Paper No. 1730, Graduate School of Business, Stanford University World Bank (1998). "A framework for civil service reform in Pakistan." Poverty Reduction and Economic Management, South Asia (December 15). 38 _____ (2001). "Nepal: Priorities and Strategies for Education Reform." Human Development Unit, South Asia Region (July 18). _____ (2002). "Poverty in Pakistan: Vulnerabilities, Social Gaps and Rural Dynamics." Poverty Reduction and Economic Management, South Asia Region (October 28). Zachariah, K. C. (1992). Demographic Transition in Kerala in the 1980s: Results of a Survey in Three Districts. Center for Development Studies, Thiruvananthapuram and Gujarat Institute of Area Planning, Ahmedabad 39 Figure 1: Public Spending on Health & Education (Per Capita 1992 Rs) 600 500 400 spending 300 capita per 200 Real 100 0 1960 1965 1970 1975 1980 1985 1990 1995 Year Kerala Uttar Pradesh Source: For nominal public expenditures, Reserve Bank of India Bulletin, various issues. Public spending data refers to spending on the current account, the largest category of state government expenditures which includes the bulk of spending on social services. Price deflator and population from World Bank India Poverty Database (Ozler et al, 1996). 40 Figure 2a: Public Spending in Kerala 0.6 Proportion of Public 0.5 Spending on Education Public & Health of 0.4 0.3 Spending 0.2 Proportion of Public 0.1 Spending on State Proportion Administration 0 (including interest payments) 1960 1965 1970 1975 1980 1985 1990 1995 Year Figure 2b: Public Spending in Uttar Pradesh 0.6 0.5 Proportion of Public Public Spending on of 0.4 Education & Health 0.3 Spending 0.2 Proportion of Public Spending on State 0.1 Proportion Administration 0 (including interest) 1960 1965 1970 1975 1980 1985 1990 1995 Year Source: Reserve Bank of India Bulletin, various issues. Public spending data refers to spending on the current account, the largest category of state government expenditures which includes the bulk of spending on social services. 41 [NB1] INSTRUCTIONS FOR FILLING PRR FORM · The data you have supplied have been entered in the proper places. Now you are in your document, ready to enter your Accomplishments. · If you haveView Field Codes checked, turn them off now. · A \ (backslash) has been entered in all the places you are required to enter text in. You can Search for a \ (backslash) to go to the cell to enter your text or just click at each \ (backslash) and type your text. · If you go back and make any changes to the text entered via macro (i.e name, title, level, staff number), remember to make the changes in the header also. · You can delete the Annotation mark (NB) in your form, when you are done reading it.