WPS5667 Policy Research Working Paper 5667 Incentive Compatible Reforms The Political Economy of Public Investments in Mongolia Zahid Hasnain The World Bank East Asia and Pacific Region Poverty Reduction & Economic Management Unit May 2011 Policy Research Working Paper 5667 Abstract Why do politicians distort public investments? And given ministry would have access to huge rents and become that public investments are poor because presumably that politically too powerful. The identity of these “winners” is what is politically rational, what types of reforms are is uncertain ex ante, given the relatively egalitarian and likely to be both efficiency improving and compatible ethnically homogenous nature of Mongolia’s society and with the interests of politicians? This paper explores these polity. Anticipating this risk, members of parliament two questions in the context of Mongolia. It argues that are reluctant to fund these projects. Since these large Mongolian members of parliament have an incentive infrastructure projects are crucial for national growth, to over-spend on smaller projects that bring benefits to neglecting them hurts all members of parliament. specific geographical localities and to under-spend on Members of parliament will therefore support reforms large infrastructure that would bring economic benefits that collectively tie their hands by safeguarding large, to Mongolia on the whole. The incentive for the former strategic investment projects from political interference is that members of parliament internalize the political thereby ensuring that no political faction becomes too benefits from the provision of particular, targeted powerful. This protection of mega-projects would need benefits to specific communities. The disincentive for the to be part of a bargain that also allows geographical latter is that large infrastructure carries a political risk targeting of some percentage of the capital budget. because the political faction in control of that particular This paper is a product of the Poverty Reduction & Economic Management Unit, East Asia and Pacific Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at zhasnain@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team INCENTIVE COMPATIBLE REFORMS: THE POLITICAL ECONOMY OF PUBLIC INVESTMENTS IN MONGOLIA Zahid Hasnain* JEL Codes: H41, H54, D72 Acknowledgements: This paper has benefitted greatly from comments by Richard Allen, Jim Brumby, Coralie Gevers, Phil Keefer, Yasuhiko Matsuda, and Rob Taliercio. * The World Bank 1 Incentive Compatible Reform: the Political Economy of Public Investments in Mongolia 1. Introduction Investments are a key to economic growth and in developing countries public investments in infrastructure, education, and health are necessary to provide public goods, address market failures, crowd in private investments, and achieve redistributive goals. However, public investments are often the most politically visible and discretionary forms of public expenditures, and are therefore strongly influenced by political considerations, thereby undermining their quality. These political pressures are particularly pronounced in resource-rich countries — as is often noted, the resource curse is primarily a political and not economic phenomenon.1 A fundamental question therefore, is how can public investments be improved in developing countries given these political constraints? More specifically, given that public investments are poor because presumably that is what is politically rational, what types of institutional and organizational reforms are likely to be both efficiency improving and compatible with the interests of these actors? This paper addresses this question in the context of Mongolia, a natural resource (copper and coal) rich and rapidly transforming economy that is presently attempting numerous structural reforms along the length of the natural resource ―value chain‖ but which also faces many of the risks associated with the resource curse. An effective public investment system is essential for avoiding the Dutch disease and for translating natural resources underground into productive assets above ground. However, Mongolia‘s public investment management system is weak and has been unable to allocate resources well or in the most cost effective manner. With the large infrastructure investment needed to both develop the new mines in the Southern Gobi (estimated at $5 billion, or roughly 100% of GDP) and to benefit from them, improving this system is a central and urgent priority. The issue of which technical solutions will be politically viable is therefore, of considerable importance to the country‘s future development. The paper‘s approach is to apply the lessons from the considerable academic political economy literature in attempting to answer this question. The key political actors in the analysis are Mongolia‘s members of parliament (MPs), the reason being that Mongolia is a parliamentary democracy in which the parliament exercises considerable budgetary authority and the public investment system is hugely influenced by the interests of individual MPs. Drawing on this literature, the paper identifies three perverse political incentives that need to be addressed if public investments are to improve in Mongolia. First, MPs have an incentive to over-spend on smaller projects that bring benefits to specific geographical localities, and which, given Mongolia‘s geography, tend to be uneconomic, and to under-spend on large infrastructure that would bring economic benefits to Mongolia on the whole. The reason is that while Mongolian MPs internalize the political benefits from the provision of particular, targeted benefits to specific communities, they have a disincentive to make the large infrastructure investments for which they will share credit with others. In particular, given the weaknesses in the public investment system, the provision of large infrastructure carries the risk that the political faction in control of that particular ministry would become politically too powerful. Anticipating this risk, MPs are reluctant to fund these projects. Second, even those projects that are a priori economic are compromised in their implementation through poor planning and appraisal, procurement irregularities, and weak monitoring. The award of contracts to 1 Humphreys, Sachs, and Stiglitz (2007), Collier (2010), and Dunning (2008). 2 technically unqualified but politically well-connected construction companies is a major political reason for these implementation problems as these contracts provide the avenue for electoral campaign finance. Third, there is little incentive for political actors to maintain capital assets as evidenced by the chronic under-funding of capital maintenance and repairs, and the resultant deterioration of physical assets. This neglect of maintenance, resulting in a ―build-neglect-rebuild‖ feature for infrastructure is a common occurrence in developing countries, and is reflective of MPs short time horizons linked with the election cycle. Mongolia shares these problems with many developing countries. However, encouragingly, there are a number of features in Mongolia that also suggest that reform is possible. First, compared to other young democracies, Mongolia is a relatively stable two-party system, and this existence of at least two parties that are capable of winning office greatly enhances electoral accountability. Second, although Mongolia‘s political parties are factionalized they are more programmatic — i.e. have some shared corporate policy objectives — and disciplined than is the case in many other developing countries. Third, there is little prospect of an extra-constitutional intervention (e.g. military coup), thereby implying that Mongolian policy-makers have longer time horizons than many of their developing country counterparts. In other words while MPs may lose elections, they are not likely to lose their life or their freedom, and have the chance of winning in the future. Fourth, Mongolia is not riven by ethno-linguistic cleavages, as is common for example in many African democracies, a major advantage given that in natural resource rich countries the combination of ethnic rivalries and resource rents has often proven to be explosive.2 Fifth, Mongolia is a relatively egalitarian society that does not have traditional, landed elites or other ―oligarchs‖ that dominate the economic and political landscape. Sixth, there is a vibrant and free press, and numerous civil society advocacy groups, that can provide the necessary checks on the executive. Finally, in large part due to the legacy of communism, the education level of the population is high compared to other countries at similar income levels, increasing the ability of the population to hold policy-makers accountable. These positive features suggest that Mongolian policy-makers should in theory be able to cooperate to achieve reforms. The key reason why reform is possible is as follows: Mongolia‘s polity currently does not have any dominant ―oligarchs‖ who combine economic and political power; instead this power is more evenly distributed among a number of members of parliament. However, the huge increases in natural resource rents that are imminent greatly increase the likelihood of certain MPs becoming oligarchs, which would imply that a minority of MPs would become politically dominant, thereby reducing the power and influence of the majority. The identity of these ―winners‖ is unknown ex ante, given that there are no presently privileged groups (relatively egalitarian polity, and no ethno-linguistic divisions). Therefore, there will be an interest among a majority of the MP‘s (the potential ―losers‖) to collectively tie their hands to prevent such as outcome. The fact that Mongolian political parties are reasonably cohesive and programmatic suggests that MPs should, in theory, be able to cooperate for this purpose. This logic implies that the following reforms are incentive compatible:  MPs will support reforms that collectively tie their hands by safeguarding large, strategic investment projects from political interference. Given the positive features of Mongolia‘s political economy, there is a recognition among MPs that they are stuck in a ―prisoner‘s dilemma‖ situation where what is individually politically rational leads to sub-optimal collective outcomes. No individual MP will find it in his interest to forgo his constituency-specific project 2 Collier (2008) 3 to free up the funding that is needed for the large infrastructure projects to develop the Southern Gobi mines. Why? Because (a) this risks hurting their re-election chances and (b) without some safeguarding of the mega-projects, a few privileged MPs — those who are able to secure these large contracts — will gain at the expense of a majority whose projects cannot get funded. The identity of these winners will be uncertain ex ante. The collective result will be that such projects will not be developed, which will hurt all MPs. Therefore, budget, planning, and procurement reforms that put in place checks and balances and collectively limit all MPs ability to influence these projects, and do not advantage or disadvantage any one particular faction will be incentive compatible.  The corollary to the above, given the strong public demand for seeing some of the benefits of natural resources, some constituency specific projects are politically necessary. Indeed, protection of mega-projects would need to be part of a bargain that also allows geographical targeting of some percentage of the capital budget. These projects will most likely be un- economic; but without them, MPs cannot get re-elected. The focus should therefore be to ensure that (a) these projects are properly prepared, in particular that their cost estimations are accurate and community consultation is institutionalized; and (b) that they cumulatively amount to a small proportion — less than 20 percent — of the annual capital budget.  The functional form for safeguarding large, strategic national infrastructure projects will also need to be incentive compatible. In Mongolia, MPs will reject an all-powerful Ministry of Finance that combines the planning and budgeting function as this will privilege one political faction (e.g. led by the Minister Finance) over others. Instead, a division of responsibilities between a central planning agency (e.g. a Ministry of Planning), the Ministry of Finance, and line ministries will be incentive compatible as it provides representation to the various political factions. The focus of reforms should then be to ensure that proper coordination mechanisms exist between the three agencies.  A similar logic of safeguarding large projects that benefit all constituencies while allowing targeted benefits to flow from smaller projects applies to procurement arrangements. International procurement standards of competitive bidding and transparency will be incentive compatible for the large mining infrastructure projects precisely because the stakes are so high and a particular political faction could become too powerful. However, for smaller projects collusive practices will be difficult to prevent, particularly given the less interest of international companies. The paper is inductive in its structure. Section 2 analyzes the broader public expenditure patterns in Mongolia in order to identify policy-makers priorities (i.e. revealed preferences), and then digs deeper into the public investment management system, identifying the main technical and structural weaknesses in the system.3 Section 3 analyzes what interests of key political actors are served by poor investments. Given these political incentives, section 4 then proposes some incentive-compatible reform options. 2. Key Issues in Public Investment Management in Mongolia Mongolia, like many revenue rich and former socialist countries, has a big government as measured by revenue collection and public expenditures. As Figure 1 shows, public expenditures have averaged 30 percent to 40 percent of GDP over the past 5 years. Also like many resource-rich countries, expenditures have been highly pro-cyclical, in large part reflecting the trend in copper prices. Mongolia has gone 3 This section draws on World Bank (2010) and Finch & Fritz (2010). 4 through a classic boom-and-bust cycle from 2005 to 2009, with the non-mining deficit (i.e., expenditures relative to non-mining revenues) increasing from a deficit of 1 percent of GDP in 2005 to a deficit of over 15 percent of GDP in 2008 (Figure 1, left panel), which required large expenditure cuts in capital and operations (goods and services), and donor assistance. Figure 1: Mongolia’s public expenditure patterns, 2003-2010 Expenditure Main government expenditures (% of GDP) Mining-related revenue 60 Non-mining revenue 8,000 40% Budget balance 35% 50 Non-mining balance 7,000 30% Copper price (end-year, right axis) 40 6,000 25% 20% 30 5,000 Percent GDP 15% 20 4,000 10% 5% 10 3,000 0% 0 2,000 2003 2004 2005 2006 2007 2008 2009 2010 2003 2004 2005 2006 2007 2008 2009 (June budget -10 amend) 1,000 Wages and salaries Subsidies Transfers Purchase of goods and services -20 - Investment Source: Ministry of Finance The use of the windfall mining revenues provide a good measure of policy-makers‘ preferences as it is like ―manna from heaven‖ that can be spent without requiring an additional tax burden on citizens (Figure 1, right panel). The 2005 to 2008 boom was used to finance massive increases capital expenditures (from 3% of GDP in 2005 to 8% of GDP in 2008), across-the-board wage increases for civil servants (from 6% of GDP to 9% of GDP), and social transfers (from 8% of GDP to 10% of GDP). Within public investments, the roads and transport, and fuel and energy sectors received a bulk of the resources (roughly half of the capital expenditures were on these two sectors), followed by education and health. The social transfers were in general untargeted and involved significant cash handouts to all Mongolian citizens with children.4 The expenditure categories that were neglected, relatively, were operations (purchase of goods and services, which declined from 11% of GDP to 8% between 2005 and 2008) and in particular capital repair and maintenance, as described in more detail below. The problem in Mongolia, unlike many other developing countries, is therefore not under-spending on capital but the poor quality of spending, as elaborated below. It should also be noted, that apart from its broader macro-stability implications — the spending effect resulting in the overheating of the economy and real exchange rate appreciation, thereby hurting the non-resource tradables sector — a fiscal cycle itself compromises investment quality. The rapid increase in expenditures — in nominal terms, public capital expenditures increased seven-fold in three years — puts a strain on the systems for planning and execution, and results in the misallocation of resources. In other words, if capital expenditures grow this rapidly then it is highly likely that they are financing many bad projects as the systems for appraising and selecting projects cannot grow in capacity and sophistication at a similarly rapid pace. Equally importantly, the local construction sector usually cannot handle this surge in demand for public projects, resulting in myriad implementation problems. Both smoothing expenditure fluctuations to avoid the problem of many new projects that then cannot be funded during the downturn, and controlling expenditure growth to preserve the quality of the capital asset portfolio is therefore imperative for a resource-dependent economy. 4 The Child Money Program 5 As is generally accepted, a good public investment system needs to do two things5:  To achieve allocative efficiency, that is to distribute resources as per national priorities, and to shift resources from less productive sectors and less effective activities to more productive and effective ones. In other words, ‗spending in the right areas‘;  To achieve operational efficiency, or to implement programs and deliver services at the lowest cost per unit of output, or conversely, to get the most output for a given set of inputs As detailed below, Mongolia‘s system is falling short on each of these objectives. 2.1. Low allocative efficiency There are two main aspects to this problem in Mongolia. First, the capital budget is fragmented with a portfolio skewed towards a large number of small projects spread across sectors and regions and an underfunding of large infrastructure investments that are critical for economic growth. Second, there is a gross neglect of maintenance and capital repair leading to the ―build-neglect-rebuild‖ scenario. Underlying these expenditure patterns are institutional weaknesses — public investment projects are not guided by well thought out strategic priorities, there is little in the form of economic appraisal in the planning and budgeting process, and capital budgeting is not done under reasonably disciplined sector ceilings and therefore there is no incentive for ministries to prioritize. The vast majority of projects in the public investment portfolio in Mongolia are small and geographically targeted which has resulted in considerable fragmentation of the capital budget into a large number of small projects. To take the example of the 2008 capital budget, the total project portfolio for new investments of roughly $900 million was spread out over 758 projects, implying an average project allocation of $1.2M. However, the median project size was much lower. As Figure 2 shows, 66% of projects were less than $400,000 in size (500 million MNT), and another 28% were under $4 million. If one considers $8 million (10 billion MNT) as the threshold for a ―large project‖, then only 2% of the portfolio was made up of these large projects. In terms of cost, these large projects accounted for 40% of the portfolio. Figure 2: Distribution of investment projects by size, This focus on smaller projects has come at the 2008 expense of a relative neglect of large, critical 2% are infrastructure investments. This is evident in the larger than inability of policy-makers to increase electricity $8M generation capacity, particularly for Ulaanbaatar, resulting in a near crisis situation of power shortages. Specifically, the need for a new Power 28% of projects are Plant # 5 for Ulaanbaatar has been on the agenda between 758 total projects for the past decade but has been repeatedly $400k and delayed. More significantly, the government has $4M in size 66% of the been unable to effectively plan for the huge projects are infrastructure investments on roads, power less than $400k in supply, and water supply, estimated at roughly $5 size billion, to effectively develop the Southern Gobi mines. Source: Ministry of Finance 5 Based on World Bank (1998) 6 Such high fragmentation also reduces allocative efficiency for other reasons. First, these projects, given Mongolia‘s geography (small population and low population density) tend to be uneconomic. Second, fragmentation likely results in duplication and wasted resources. Third, it likely increases the costs of the public investment portfolio as it forgoes some economies of scale. Fourth, the large number of small projects taxes the limited human resource capacity of the line ministries that prepare and implement these projects, and of the Ministry of Finance that has to screen and approve them to be financed from the budget. It is therefore likely that a portfolio with many smaller projects will be of poorer quality than a portfolio of the same size with fewer, larger projects. Another stark objective indicator of the misallocation of resources in Mongolia is the gross under- spending on capital maintenance and repair in relation to the country‘s needs. Due to years of neglect, the state of disrepair in the energy and roads sectors is approaching crisis proportions. The electricity sector has immediate capital repair needs estimated by the Ministry of Fuel and Energy of about $130 million (or 2.5 percent of GDP), and 60% of the national paved road network is in poor condition and in need of capital repair and rehabilitation, at an estimated cost of $520 million (or roughly 10 percent of GDP). As against these large needs, and as shown in Figure 3, left panel, capital repair spending increased very little in nominal terms during the boom years, as compared to expenditures on transfers, wages and salaries, and new domestic investment, and declined as a percentage of new investments (from 10 percent to 6.5 percent) and of estimated public capital stock (from 4 percent to below 2 percent) between 2003 and 2010.6 Figure 3: Spending in capital repairs and maintenance did not benefit from the expenditure boom and has declined as a share of new investments and capital stock Main government expenditures Trends in capital repairs 12 3,000 Capital repair 2,500 10 Billions 2,000 8 1,500 As percent of new 1,000 6 investments 500 As percent of public 4 capital stock 0 2003 2004 2005 2006 2007 2008 2009 2010 2 budget Wages and salaries Subsidies and transfers 0 Purchase of goods and services Doemstic investment 2003 2004 2005 2006 2007 2008 2009 2010 bud. Capital repair Source: Ministry of Finance, World Bank staff calculations. Public capital stock calculated using perpetual inventory methodology specified in Harberger (2007). There are both technical and political reasons for this under-prioritization. As elaborated on below, there is no economic analysis of projects during budgeting and no requirement that the future recurrent costs of projects be specified and be part of the appraisal process. The deeper political reasons are discussed in the next section. Ideally, budgets generally, and public investments specifically, need to be guided by policies that are grounded in a strategic framework based on sound analysis. The requirement therefore is both that national and sectoral planning documents be of good quality, and that the public investment program be, as much as possible, linked to the priorities as established in these documents. While there are numerous 6 The Government does not have an asset register and therefore capital stock is not measured. For the purposes of this report the public capital stock was estimated using the methodology in Harberger (2007). 7 national and sectoral planning documents in Mongolia, these are usually quite weak in quality, and there is very little linkage between the documents and the annual capital budget. Exacerbating this problem of weak strategic guidance is the failure to effectively enforce sector ceilings on line ministry budget proposals with the result that there is little incentive for the ministries to prioritize their investment proposals given fiscal constraints. For example, in 2008 the investment proposals submitted by all line ministries amounted to 22.6 percent of GDP, as against a ceiling of 8 percent of GDP set out in the medium term fiscal framework.7 Politically, this lack of discipline implies that ministries are not forced to make tough choices and instead can ―pass the buck‖ to the Ministry of Finance by including numerous projects and thereby pleasing all constituencies. There is no formally institutionalized process of economic appraisal of investment projects in Mongolia. The primary budget legislation (the Public Sector Finance and Management Law of 2003) is largely silent on the modalities for public investment planning and budgeting, and the Ministry of Finance regulations on preparation and prioritization of public investment projects are highly skeletal. There are no project appraisal guidelines that specify what economic and financial analysis needs to accompany project proposals, and what will be the basis of appraising these projects. Without such systematic approaches, it is very difficult to determine inter-sectoral and intra-sectoral priorities, and to determine which projects are feasible and which are not. The corollary to this regulatory weakness is the lack of capacity in the line ministries to conduct economic analysis, and the lack of capacity in the Ministry of Finance for an independent review of project proposals based on analytical criteria.8 The Investment Division of the Ministry, which until the recent creation of the new planning agency, had prime responsibility for this task has a staff of strength of less than 10, and for all practical purposes the project review process is limited to ensuring that the proposals for new constructions have the necessary technical drawings to comply with the requirements of the Construction Law, and that the cost estimates follow the required norms. These problems are particularly severe in projects Figure 4: Parliament has significantly expanded the that are inserted by parliament during the annual capital budget budget sessions. Parliament has regularly significantly increased and altered the capital Capital Budget (Public Investment Plans), Billions of budgets submitted by the executive — In 2008 MNT 571.4 for example, the parliament increased the capital 518.8 531.4 Executive budget approved by the cabinet from 442 billion 456.1 452.5 447.8 Parliament 427.4 442.7 MNT to 571 billion MNT, replacing and adding Amended numerous projects to the portfolio (Figure 4). 364 Some of these new projects flow from the Local 303.8 Development Fund, or the allocation given to each parliamentarian for earmarked or ―pork- 169.8 barrel‖ spending in their home districts. These 137.5 constituency funds were introduced in 2003, and 69 the allocation per MP has increased hundred-fold since then from 10 million MNT per MP in 2003 2006 2007 2008 2009 2010 to 1 billion MNT per MP in 2010. Often these Source: Ministry of Finance schemes identified under the constituency fund are without basic technical documentation and have even more inaccurate cost estimations than the projects proposed by the line ministries. 7 Finch & Fritz (2010). 8 See Rajaram et al (2008) for the framework for key aspects of a good public investment management system. 8 The Local Development Fund amounts to approximately 15% of the capital budget. However, the total number of constituency-specific projects is much larger, as many projects proposed by line ministries also have this feature. As noted above, 66% of projects are small and often geographically targeted, a more accurate measure of the scale of constituency-specific schemes. 2.2. Low operational efficiency These problems in allocative efficiency are compounded by problems in the project cycle. Delays in project completion are common and significant. 65 percent of roads projects completed between 2005 and 2009 had time overruns, more than doubling the original completion time in some cases, while 73 percent of on-going projects have been extended. These delays reflect two factors: poor planning and inaccurate cost estimation, and weaknesses in the construction sector. Roughly 32% of these projects were delayed due to upfront problems of planning, 49% were due to the downstream problems in implementation due to the weak capacity of the construction companies, and 19% of the projects were delayed due to both reasons. Inaccurate cost estimation usually translates into time over-runs primarily because in Mongolia the capital budget is approved and appropriated project-by-project by the parliament and increasing the budgeted cost of a project requires a budget amendment. There are three main reasons for the poor cost estimation: use of overly simple methods to estimate costs that served as the basis for project appropriations, projects inserted into the budget by parliament that were not subjected to cost estimation to the same degree as executive-proposed capital projects, and increases in building material prices, particularly during the inflationary period of 2008-09. Given Mongolia‘s short construction season (from May to October), the initial months after the capital budget is approved by parliament in December are critical for timely implementation of the project. Procurement however is a major bottleneck in the process, due to both unrealistic cost estimates during project preparation and lack of procurement capacity in the executing agencies. As the capital budget has the force of law, and includes both the overall cost of the project and allocation for the particular fiscal year, no bid can be awarded that is above the approved cost of a project. Inaccurate cost estimates during project preparation imply that on many occasions the bids received do not fall within these cost ceilings, and on average procurement can take 6 months. In the last two years this problem has been compounded by high inflation that has resulted in significant cost escalation during implementation. In many cases line ministries have had to seek parliamentary amendments to the capital budget to increase the cost allocation of projects to make them more in line with market rates. This problem also encourages construction companies‘ compromise on quality in order to avoid incurring losses.9 These procurement delays are particularly severe for the projects that are added by the parliament as they usually have no accompanying engineering designs. Weaknesses in the legal framework for procurement also raise questions on the quality of capital budget execution. Procurement is decentralized in Mongolia to the line ministries and agencies, with the Procurement Policy and Coordination Department of the Ministry of Finance setting policy and providing the regulatory and monitoring function, and conducting prior reviews for large contracts. As a recent Government‘s audit report details, the major problem in this framework is the provision for direct contracting that was introduced in 2007 in the roads and energy sector. In 2007, 125 billion MNT, or 34 percent of all of contracts, were awarded through direct contracting, often with no accompanying technical documentation with allegedly numerous technically unqualified, and politically well connected, companies winning construction contracts (Government of Mongolia, 2008). Many road construction 9 In 2008 there was an exception due to significant inflationary pressures on raw material costs. Contractors were awarded MNT 4 billion for cost overruns. 9 projects were also broken down into smaller schemes to enable smaller and less qualified companies to be directly awarded the contract. This practice of direct contracting was discontinued in 2008 and 2009, but remains allowable in the Procurement Law. Figure 5: Time overruns in the ten sampled projects Plan project time Overrun Overall sample average 1.8 1.3 Directly contracted projects Plan Overrun Competitively tendered projects 0 1 2 3 4 5 Years Source: Department of Roads The findings from case studies of a sample of road sector projects corroborate the findings of the audit report. Of the total sample of ten, four projects were executed through direct contracting while the remainder went through the normal tendering process. As portrayed in Figure 5, seven of the ten projects had time overruns, with an average completion time of 3.1 years compared to the original plan time of 1.8 years, or an overrun of 70 percent. The directly contracted projects were the worst performers, with an average time overrun of 2.25 years (or 180 percent). Other reasons for the time overruns are bottlenecks in financing and the lack of implementation capacity of construction companies. The Ministry of Finance exercises strict input controls in the execution of the budget; the capital budget is approved project-by-project as law by parliament annually, even for multi- year projects, there is no multiyear appropriation, nor any provisions for carry-over funding. Line ministries have no authority to reallocate funds across projects based on revised cost estimates or pace of implementation, and any unspent allocation gets re-appropriated to the Ministry of Finance at the end of the fiscal year. For multi-year projects there is no prioritization process that would ensure that projects that are being implemented well would receive their necessary annual allocation. 3. The Politics of Public Expenditures The assumption behind this paper is that these technical weaknesses to a considerable extent reflect political choices, and that several perverse incentives need to be addressed if public investments are to improve in Mongolia. Clearly there are a number of important political actors in this process. This paper focuses on members of parliament. The reason for this focus is both that the Mongolian parliament has considerable formal budgetary authority and that individual MPs exercise significant informal influence over the executive in project selection and execution. Mongolia is a parliamentary democracy based on a majoritarian electoral system with elections held regularly every four years since the transition from socialism in 1992. The two largest political parties are the Mongolian People‘s Revolutionary Party (MPRP) and the Democratic Party; the MPRP, which was 10 the old communist party, has won clear majorities in three elections (1992, 2000, and 2008). The parties are not particularly distinguished by ideology — for example, both have campaigned on broadly populist platforms of redistributing mining wealth to the people in the form of large, untargeted cash transfer programs, and of investment projects distributed across the country — and both are characterized by factionalism. Figure 6: The formal budgetary authority of Mongolia’s parliament stands out in comparative perspective Index of Formal Legislative Authority United States Sweden Norway Hungary Netherlands Austria Mongolia Argentina Germany Denmark Belgium Upper quartile: 66.7 Uruguay Turkey Spain Portugal Mexico Japan Indonesia Iceland Finland Median: 55.7 Czech Republic Bolivia Slovenia South Korea New Zealand Italy Suriname Slovakia United Kingdom Lower quartile: 33.3 Israel Greece Canada Australia Ireland France Chile South Africa 0.0 20.0 40.0 60.0 80.0 100.0 Source: Author‘s calculations based on Wehner (2006). The index is a sum of three indicators normalized to range between 0 and 100: legislative powers of amendment; consequences of non-approval of the budget; and executive flexibility on budget appropriations The parliament in Mongolia exercises unusual budgetary powers. As shown in Figure 6, Mongolia‘s parliament ranks in the top quintile in formal budgetary authority (powers of budget amendment, consequences of non-approval of the budget, and executive flexibility in budgetary appropriations). As noted above, this formal authority has been repeatedly used by MPs to amend the budget in general and the capital budget in particular. Equally important is the informal political interference in the execution of the overall capital budget through in particular the influencing the contracting process. As discussed above, the most egregious example of this influence was the large number of contracts awarded without competitive bidding in 2007. Anecdotal evidence suggests that a prime motivation for these contracts was to generate political support from wealthy and influential individuals ahead of the national elections in 2008. In general, political interference in the award of contracts is known to be considerable even when competitive bidding is done de jure, in particular given the small size of the formal sector in Mongolia, and the close personal links between parliamentarians and the business community. 11 These incentives of elected officials are a common feature of developing country democracies. The sub- sections below examine the relevant academic literature on political economic resource allocation in democratic settings and then apply the lessons from this literature to the Mongolian context. 3.1 Lessons from the literature There are three relevant academic literatures to understand the politics of public investments in Mongolia. The first is the literature that has focused on the political economy of macro-economic policy, in particular the incentives that result in fiscal indiscipline. The second is the literature that has attempted to explain why democratic governments that are ostensibly accountable to voters would systematically under-provide public goods thereby hurting the welfare of these voters. And third, is the huge literature on the political economy of reform. Fiscal indiscipline is usually explained in the political economy literature as a common pool resource dilemma in which the political and institutional structure of the country allows for decentralized spending determination. Specifically, policy-makers have the ability to target spending to particular constituencies using general tax revenues that are shared across all the constituencies. Under these conditions, politicians have the incentive to overspend as they internalize the full benefit of the local expenditures but only bear a fraction of the increased costs associated with these expenditures.10 This negative externality generates over-spending and fiscal indiscipline. These problems are likely to be exacerbated in natural resource-rich environments in which the revenues flow from resource rents rather than taxes. The literature suggests that both centralizing fiscal authority and cooperative bargaining are conducive to overcome the inefficiency and thus are able to promote fiscal discipline. Several empirical studies have shown that budget institutions have an impact on fiscal outcomes in a variety of samples of developed and developing countries.11 Market failures are also fundamental to explaining why democracies can under-provide public goods that help the majority of the population (and voters). The huge literature on clientelism attempts to explain this puzzle by the relative inability of politicians to internalize the benefits of public as opposed to private good provision. These targeted goods can be ―pork-barrel‖ projects, or projects that benefit specific localities, or ―rents‖ distributed to influential groups in exchange for their political support. Bates (1981) classic formulation of clientelism argued that economic interests that can organize collectively can provide politicians with political support, which creates an incentive for politicians to provide rents to these groups. Society suffers because the particular interests of these groups are inimical to broader development. Cross-country variations in development could then be attributed to either the strength and number of interest groups (Olson 1984), with for example in pluralistic societies multiple interest groups competing for state attention cancelling each other out (North, Wallis, Weingast 2006); or whether or not these special interests were antagonistic to economic development (e.g. Acemoglu and Johnson 2002); or on the whether or not economic activity is concentrated or dispersed (Dunning 2008). More recent work has shown that politicians will have an incentive to provide targeted benefits to their supporters in situations where they will not receive electoral benefits from providing public goods either due to the inability to mobilize support for public goods across ethnic groups (Easterly & Levine 1997, Alesina et al 2005, Keefer 2010), voters lack of information about the actions of individual politicians (Besley, 2006), or lack of credibility of national political organizations for providing these goods (Keefer 2007). 10 Weingast et al (1981), Velasco (2000) 11 Alesina et al. (1996). 12 Information can also be a key reason as to why democracies can produce outcomes that hurt the median voter. Politicians care about getting elected, and they will focus on policies and will emphasize expenditure in areas that will maximize their electoral fortunes. Importantly, getting elected is also crucially contingent on information – specifically, it requires that voters are able to connect improvements or deteriorations in their welfare to the actions of a particular politician. In other words, politicians must be able to credibly take credit for these improvements. This importance of information implies that politicians are likely to be more responsive to the more informed group of citizens, and as a result focus more on private goods, or targeted benefits, as opposed to public goods that benefit the majority.12 Alternatively, even if voters are informed they may not consider as credible politician‘s promises to improve public goods. A credible politician is one that can provide targeted goods to those citizens he develops a personal interaction with, which are typically influential people from his constituency. This incentive for targeted goods is also driven by the fact that elections are expensive and a politician usually has to raise funds himself to finance his campaign. The exchange of rents for campaign contributions as well as votes is then another incentive for the under-provision of public goods and the over-provision of targeted goods. A number of institutional variables may impact clientelism. There is a large literature that has explored the impact of regime type (presidential vs. parliamentary) and electoral rules (majoritarian vs. proportional) on targeted spending. No conclusive results have emerged, primarily due to the considerable heterogeneity of institutional arrangements within the above stylized categories. There is an emerging literature on the influence of political parties on clientelism which suggests that more cohesive political parties, and stronger voter attachment to political parties — or partisan voting — reduces the incentives towards clientelism and encourages a more national, programmatic orientation in individual members of parliament.13 There are numerous refinements to the above argument. Robinson and Verdier (2003), and Robinson and Torvik(2005) ask the question why clientelism should necessarily be so inefficient, manifesting either in the form of public sector employment or ―white elephant‖ projects with negative social return. The answer is that clientelism is inefficient precisely because this inefficiency solves the commitment problem (i.e. why would a voter honor his ‗contract‘ by voting for the politician after receiving the private good?): only by maintaining that particular politician responsible for the rents in office can clients hope to keep receiving the rents. If the rents were efficient, then all politicians would have an incentive to keep supplying them and voters would have no incentive to vote in a particular politician. To summarize this literature, public goods are under-supplied and private goods over-supplied because (a) rents can be exchanged for campaign contributions; (b) rents can be targeted which solves the information problem and the attribution problem (the politician can take credit for them); (c) the provision of rents implies that a politician is credible; and (d) rents will only continue if voters re-elect that particular politician, thereby solving the commitment problem. Finally, there is the huge literature on the political economy of economic and political reform. Here, we focus on two works of particular relevance to the argument of this paper. One, by Fernandez and Rodrik (1991) shows how the combination of individual specific uncertainty concerning whether you will win or lose from reform, together with majority rule, may cause socially beneficial reforms to be blocked even though everybody knows that a majority will gain from reform ex post. The logic rests on the fact that 12 Important recent works in this literature include Keefer (2002); Keefer and Khemani (2003) 13 Keefer and Khemani (2003) 13 the identity of the winners and losers is uncertain ex ante, and the winners cannot credibly commit to compensate the losers ex post. A second, by Acemoglu and Robinson (2006) explains the puzzle of why elites may block growth enhancing reforms by the ―political replacement effect‖ — i.e., while reforms may increase growth they may also foster changes that result in the elites being removed from power; anticipating this, the elites will block reforms. The relationship between blocking reforms and political competition is non-monotonic. Both political elites that are subject to competition and those that are highly entrenched are likely to reform. With intense political competition, elites prefer to innovate, because otherwise they are likely to be replaced. With a high degree of entrenchment, incumbents are willing to innovate, because they are not afraid of losing political power. Instead, it is elites that are somewhat entrenched but still fear replacement that will block innovation. 3.2. Applying the lessons to Mongolia Given the numerous socio-economic and institutional variables identified in the literature, and the public expenditure patterns and problems in the public investment system discussed above, where does Mongolia stack-up in terms of policy-makers incentives for providing sound public investments? The following negative and positive aspects are suggestive: The negative aspects:  The combination of high parliamentary spending authority, large natural resource rents, the majoritarian electoral system with MPs representing specific geographical constituencies, and need for constituency-specific spending for re-election purposes create a classic common pool problem that encourages over-spending by legislators and is the main political cause for the boom-and-bust cycle.  The fragmentation of the capital budget — large number of small, often uneconomic projects, targeted to specific geographical areas — can be explained by the need for individual members of parliament to cultivate their own, individual-specific political support in order to win elections.  The relative neglect of large, nationally important infrastructure is suggestive of the inability of parliamentarians to work towards cooperate goals. The main reason for this is that given that institutional arrangements for public investments are weak large infrastructure projects carry a particular risk of political capture that can threaten to create a dominant political faction given the high potential rents involved. Following Fernandez and Rodrik (1991) the identity of these winners is uncertain ex ante; and given this, following Acemoglu and Robinson (2006), the MPs have an incentive to block such large projects.  The election cycle encourages short time horizons in MPs which creates the lack of interest in maintaining capital assets, resulting in the ―build-neglect-re-build‖ project scenario. .  Procurement irregularities can similarly be explained by clientelist pressures; MPs trade construction contracts for campaign finance and political support. The positive aspects:  Even though Mongolia‘s political parties are factionalized, they are more programmatic than many in other developing country democracies, as is detailed below. Programmatic parties are those in which members share common policy preferences and where citizens hold 14 politicians from the party collectively responsible for failure to pursue policies that are consistent with the party‘s program. The untargeted social transfer programs of both the MPRP and the Democratic Party are indicative of programmatic elements.  Even though individual MP‘s face considerable uncertainty regarding their re-election prospects, which encourages them to focus on ―consuming now‖ rather than ―investing for the future‖, Mongolia is a relatively stable democratic system with little prospect for a regime change. Therefore, there is some assurance for individual MPs that while they may lose the next election, they can eventually return to power, and will not risk getting jailed through a military coup.  Mongolia has some socio-economic advantages that reduce clientelist pressures. First is its ethnic homogeneity and strong sense of national identity. Second, is its high and growing level of urbanization (Ulaanbaatar has 40% of the country‘s population) that will reduce incentives for pork-barrel funding to rural communities. Third, is the high education level (universal literacy and over 30% tertiary enrollment rates) that should ameliorate some of the informational roots of clientelism. Fourth, is the relative egalitarian nature of the society and the absence of landed elites and powerful oligarchs. In particular, reform is possible because Figure 7: Mongolian MPs do not switch parties Mongolian political parties, while factionalized, do have a corporate identity that sets them apart Percentage of MPRP elected MPs who were past from political parties in many developing MPs from same or different political party countries. There are numerous indicators of the 80% programmatic nature of these parties. First, there is reasonably strong partisan identification among the electorate because of the strong grass-roots 42% presence of the MPRP, which as the former 29% communist party, has existed for over 80 years. Second, there has been virtually no switching of 0 0 0 MPs from one party to another — a common feature in many low and middle income Same Different Same Different Same Different democracies — suggesting a reasonable degree of 1996 2000 2004 corporate identify. As shown in Figure 7, Source: Election Commission between 1996 and 2004, there were no MPRP MP‘s who had in the past been parliamentarians from the Democratic party. Similarly there were virtually no Democratic party MPs who in the past were with the MPRP. These results are particularly striking given that the Democratic Party is a significant political force (it has been in power twice, between 1996 and 2000, and between 2004 and 2008 in a coalition with the MPRP) and therefore would certainly attract the interest of MPRP MPs. Third, there are virtually no independent candidates who win elections. Fourth, the size of Mongolia‘s cabinet has remained modest, at between 11 and 15, over the past two decades, even during coalition governments of 1996 and 2004, suggesting that MPs allegiance was not required to be bought with a cushy cabinet post to prevent defection. There are also institutional rules with parliament that help collective action by MPs. In particular parliamentary standing committees, such as those on the budget, social policy, and state structure, exercise considerable power, and their membership is based on seniority. While budget amendment sessions are essentially a time of intense bargaining as individual MPs lobby to get their pet projects into the budget, the process carried out under the authority of the budget standing committee, with the chairman of the committee exercising considerable authority. It is therefore important for individual MPs 15 to have good standing with the parliamentary party hierarchy in order to get their projects included, which encourages discipline. 4. Incentive Compatible Reforms 4.1 Fiscal responsibility legislation The ability of MPs to cooperate to address these problems was demonstrated with the passage in 2010 of the Fiscal Stability Law (FSL) that mandates that the annual budget be based on fiscal rules that limit expenditure volatility, make expenditures countercyclical, and control expenditure growth. While it is too early to tell whether the FSL will be adequately implemented, discussions with MPs revealed that the law was passed because the MPs appreciated that they were stuck in a ‗prisoner‘s dilemma‘ type of problem: that while it was individually politically rational for each MP to over-spend for the benefit of his constituency, as per the ―common pool‖ problem identified in the literature, this would lead to fiscal crises and macroeconomic instability that would be collectively irrational. The key was that the FSL collectively tied all MPs hands thereby not conferring any advantage or disadvantage to any particular constituency or faction. It was therefore incentive compatible. The fact that MPs could agree on this legislation is suggestive of the positive features of Mongolia‘s political economy listed above that allowed for a cooperative solution to this problem. It also suggests that similar reforms that collectively tie MPs hands are incentive compatible and will be the key to improving the public investment management system. 4.2. Institutional checks on large and strategically important infrastructure projects As noted, Mongolia‘s capital budget is very fragmented with numerous small projects and very few large projects. The main political reason for this fragmentation is clientelism: numerous smaller projects allow each MP to be able to target projects to his constituency as well as to exchange rents in the form of contracts for campaign financing and political support. However, the need to develop the Southern Gobi mines implies that considerable budgetary funding has to be shifted from the smaller projects to the large infrastructure ―mega projects‖. In other words the composition of the public investment portfolio will have to shift from large number of small projects, to fewer, and on average larger projects. This change in the public investment portfolio has serious political ramifications for individual MPs as (a) it risks hurting their re-election chances as constituency-specific projects are cut and (b) without some safeguarding of the mega-projects, a few privileged MPs — those who are able to secure these large contracts — will gain at the expense of a majority whose projects cannot be funded. These privileged factions will then become politically dominant; i.e. they will become oligarchs.14 Given the fluid nature of political alliances in Mongolia, helped by the fact that there are no deep and entrenched ethnic or tribal cleavages, it is likely that the identity of the winners is uncertain ex ante and the winners cannot credibly commit to compensate the losers ex post.15 Each individual MP will therefore have an interest in opposing the cutting of his pet project to fund the mega projects; the collective result will be that such projects will not be developed, which will hurt all MPs. Following the logic of the passage of the FSL MPs should have an interest in institutional mechanisms that protects these large projects from political interference, thereby collectively tying their hands and solving this ―prisoner‘s dilemma‖. Given Mongolia‘s positive stylized facts of ethnic homogeneity and regime stability, each MP has a relatively strong interest in the development of the Southern Gobi mines 14 Following the logic of the resource curse literature. 15 Following the logic of Fernandez and Rodrik (1991). 16 and therefore of increasing the size of the ―economic pie‖, as compared to, for example, their counterparts in many African countries where distributional, zero-sum conflicts among rival groups of elites, often along ethnic lines, predominate. MPs should have an interest in ensuring that the necessary infrastructure is built to enable the minerals to be produced and exported. This would imply that reforms to improve project appraisal and to protect it from political influence should be incentive compatible. These reforms, following Rajaram et al (2010) imply an institution that serves as a ―gatekeeper‘ to ensure that only projects that gone through some form of project appraisal based on cost-benefit analysis are approved and funded. 4.3. Constituency specific projects should be “formalized” Given the above shift in the public investment portfolio, it is likely that the need for geographically targeted projects will only increase in the future as (a) constituency demands to see the results of the mineral revenues will get stronger with the development of the mines; and (b) a smaller portion of the capital budget will be discretionary, given the institutional checks on the mega projects. Each MP will have an incentive to exercise control over these smaller, pork-barrel schemes so as to capture all the political benefits from them. Therefore, any reforms that try to either eliminate the Local Development Fund, or to prevent MPs from exercising influence over these constituency schemes will not be incentive compatible and will likely fail. Instead, a better reform strategy would be to take these incentives of politicians as given and to try and improve the quality of these geographically targeted projects. In other words, local development objectives should be formalized and made explicit. This formalization can be done by explicitly including the Local Development Fund in the budget preparation process so that line ministries are responsible for preparing these schemes, as opposed to the current practice in which these schemes are introduced during the parliamentary budget debates, thereby not allowing for sufficient preparatory time. Giving line ministries responsibility early in the budget cycle should improve the cost estimation of these schemes. Some aggregate limits — e.g. less than 20 percent of the capital budget — on the amount of annual funding for these schemes could also be introduced, as part of this bargain. Consultation mechanisms with local level governments and local communities in identifying these projects could also be institutionalized. Given Mongolia‘s low population density, these schemes will likely remain uneconomic from a strict cost-benefit calculus. Formalizing them will serve to at least improve the cost side of the equation; and to reduce the incentive to capture rents from the large projects that are essential for Mongolia‘s development. 4.4. Institutional arrangements for project appraisal: Function vs. form The above suggests that budget and planning legislation that specifies that all projects of national importance above a certain size threshold would need to go through a reasonably rigorous and safeguarded appraisal process would find support among MPs, provided that it is credible and specified in the legal framework. The next question then is what form would such an institutional arrangement for a ―gatekeeper‖ function take? The standard textbook advice is that planning and finance functions should be unified in a powerful Ministry of Finance which would serve as the gatekeeper to appraise and budget public investment projects. To quote Weber (2007): Separation of responsibility for current and development expenditures between two ministries – typically the ministry of finance and the ministry of planning (or economy) respectively – is often a major obstacle 17 to effective budget integration… Internationally, there is increasing recognition and acceptance of the principle that budget planning, presentation, management and accounting and control functions can be conducted more effectively within one central agency: typically a ministry of finance. The received wisdom therefore is that separating the planning and finance function leads to dual budgeting and therefore should be avoided, and forms the standard policy advice of the International Monetary Fund. Political economy factors however, suggest that such a powerful Ministry of Finance is not compatible with MPs incentives for precisely the reasons above — it would (a) risk creating a privileged political faction (the one that controls the MoF) and (b) also risk MPs‘ discretion over smaller projects. Currently Mongolia‘s Ministry of Finance has a staff of roughly 150, which has been inadequate to enable it in particular to perform the policy analysis, planning, and budget preparation function. However, proposals to increase the Ministry‘s staffing have been repeatedly rejected by parliament, in large part because all central ministries are small and increasing the MoF‘s staff would create a political imbalance and would result in similar demands from all other ministries, demands that would have backing from the different political factions represented by these ministries. Instead, Mongolia has recently created a new planning agency, the National Development and Innovation Committee (NDIC) under the prime minister‘s office. Debates amongst policy-makers preceding the creation of the NDIC included the option of strengthening the planning function in the MoF, but this option was rejected due to fears that it would result in a too powerful Ministry. While the creation of the NDIC risks fragmenting planning from budgeting and there are risks that without proper coordination between the agency and the MoF, and proper distribution of responsibilities, a dual budget system will develop, separating the planning and budgeting function is incentive compatible. And it is indeed possible that if effective technical coordination does develop between the two, then the separation of the planning and budgeting function would create the necessary checks and balances to prevent the ―capture‖ of one Ministry by political forces. This would ―divide and conquer‖ strategy may be especially effective if the Ministers of the planning agency (in this case the prime minister) and the Minister of Finance are from opposing political parties. In this case collusion would be required to ensure that a bad project gets included in, say, a multi-year public investment plan and then also gets funded in the annual capital budget. The incentive compatible reform would therefore be to give a central planning agency sufficient authority to act as the gatekeeper for large, strategically important projects, and then to ensure that the coordination between the MoF and the planning agency is strong. These mechanisms would need to be enshrined in the primary planning and budget legislation. A second institutional form question concerns the division of responsibilities between line ministries and the central planning agency in project appraisal for these large projects. Again, the logic of incentive compatible reform would imply that these responsibilities would need to be shared as the central planning agency will also never have enough staff to appraise all projects. Therefore, the NDIC would have to be limited to only large and strategically important projects, while line ministries would need to be responsible for appraising smaller projects. 4.5. Procurement reform A similar logic of safeguarding large projects that benefit all constituencies while allowing targeted benefits to flow from smaller projects applies to procurement arrangements. International procurement standards of competitive bidding and transparency will be incentive compatible for the large mining infrastructure projects precisely because the stakes are so high and a particular political faction could 18 become too powerful. However, for smaller projects collusive practices will be difficult to prevent, particularly given the less interest of international companies. This logic helps us understand the current thrust of procurement reform in Mongolia which is to centralize the procurement of large projects while giving local governments more authority over procurement arrangements for smaller projects. Presently line ministries do the procurement for all projects in their portfolio, with the Ministry of Finance responsible for setting procurement policies and monitoring, including providing no objections on projects above certain thresholds. However, numerous MPs have proposed that authority over large procurements should be taken away from line ministries and given to a newly established central tenders board, either under the Ministry of Finance or the State Property Committee, while local governments should be given procurement authority over smaller projects. The belief is that a centralized tenders board can be more easily safeguarded than a line ministry; and conversely a local government will be more susceptible to influence than a ministry. It should be noted that there is no consensus on whether or not centralizing procurement will safeguard it. On the one hand, if sufficient independence is given to the procurement authority, with strong monitoring mechanisms in place, the risk of irregularities can be reduced. On the other hand, centralizing authority over large contracts can also make the agency an attractive target. The above logic implies that ensuring that the central tenders board has the necessary institutional safeguards would be incentive compatible, and therefore should be a focus of reform efforts. 4.6 Parliamentary procedural reforms Reforms to parliamentary procedures that rationalize the budget amendment process are another potential area of incentive compatible reforms. Currently there are few restrictions on individual MPs in Mongolia on introducing budget amendments during the budget sessions, with the result that numerous additional public investment schemes are introduced greatly changing the nature of the capital budget. While the FSL will impose aggregate limits on these amendments, so that the budget in the aggregate must comply with fiscal rules, it will not impact the composition of the budget. In addition to safeguarding the appraisal and procurement of large projects, MPs will also have an interest in ensuring that the funding for these projects does not get chipped away by the introduction of numerous smaller pork-barrel schemes through budget amendments. One method would be to change the Law on Parliamentary Procedures by restricting the individual MP‘s ability to introduce amendments. In Brazil for example, another country where individual MP‘s make considerable amendments to the budget resulting in over-spending, a fiscal crises in 1993 prompted procedural changes that limited the ability of individual legislators to propose amendments, and instead gave priority to collective amendments.16 Following the Brazil example, Mongolia could introduce procedural changes that for example (a) limit an individual MP‘s ability to introduce amendments by limiting these amendments to only to small projects and (b) require collective amendments — support from a majority or even two-thirds of MPs — to amend large, strategically important projects. These separate domains for individual and collective amendments follow the logic of allowing MPs discretion over small, constituency-specific schemes while protecting the large, nationally important ones. 5. Conclusion 16 Alston et al (2005). 19 There is an emerging recognition in the international financial institutions that standard technical solutions for addressing public financial and sector management weaknesses in developing countries that are not sensitive to local ―context‖ have not worked and are unlikely to work. The need for political economy analysis is repeatedly emphasized; but there is also considerable skepticism about whether political economy work can provide a different and better, ―second best‖ solution. In general this skepticism is valid — the academic political economy literature is excellent for showing why particular pathologies exist, but less suited to providing policy guidance for some of the questions that confront institutions like the World Bank. It is hoped that this paper has provided an example of how this political economy literature can be used to develop specific and more workable policy recommendations. This literature provides a number of conceptual frameworks that can be used to analyze a particular situation in a particular context, and for testing whether potential technical solutions would be compatible with the interests of key policy-makers. The agenda for this policy relevant political economy research is wide open. 20 REFERENCES Acemoglu, Daron and James Robinson. 2006. Economic Backwardness in Political Perspective. American Political Science Review. Vol 100, No.1, 115-131. Alesina, A, R. Baqir and W. Easterly. 1999. Public Goods and Ethnic Divisions. Quarterly Journal of Economics 119(2): 565-611. Alesina, Alberto, Ricardo Hausmann, Rudolf Hommes, and Ernesto Stein. 1996. Budget Institutions and Fiscal Performance in Latin America. NBER Working Paper no. 5586. Cambridge, Mass.: National Bureau of Economic Research. Alston, Lee, Marcus Melo, Bernardo Mueller, and Carlos Pereira. 2005. Who Decides on Public Expenditures? A Political Economy Analysis of the Budget Process: The Case of Brazil. Washington D.C: Inter-American Development Bank. Bates, R. 1981. Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley: University of California Press Collier, Paul. 2010. The Plundered Planet. Oxford: Oxford University Press ----------2008. The Bottom Billion. Oxford: Oxford University Press Dunning, Thad. 2008. Crude Democracy: Natural Resource Wealth and Political Regimes. Cambridge: Cambridge University Press. Easterly, William and Levine. Fernandez, Raquel & Dani Rodrik. 1991. Resistance to Reform: Status Quo Bias in the Presence of Individual Specific Uncertainty. American Economic Review. Vol 81, 1146-55 Finch, Christopher & Verena Fritz. 2010. ―Challenges of Good Public Investment Management in a Resource-Rich Country: A Political Economy Perspective on Mongolia‘s Recent Experience. Government of Mongolia. 2010. Audit Report on 2010 Public Investment Budget. Harberger, Arnold. 2007. ―On Growth, Investment, Capital, and the Rate of Return‖. Mimeo, Los Angeles: University of California, Los Angeles Humphreys, Macarten, Jeffrey Sachs, & Joseph Stiglitz eds. 2007. Escaping the Resource Curse. New York: Columbia University Press. Keefer, P. 2010. ‗The Ethnicity Distraction: Political Credibility and Partisan Preferences in Africa‘, Policy Research Working Paper Series, 5236. The World Bank. Keefer, P. 2007. Clientelism, Credibility, and the Policy Choices of Young Democracies. American Journal of Political Science 51:4, 804-821 North, Douglass, John Wallis, & Barry Weingast. 2006. A Conceptual Framework for Interpreting Recorded Human History. NBER Working Paper no. 12795. Cambridge, Mass: National Bureau of Economic Research. 21 Olson, Mancur. 1984. The Rise and Decline of Nations. New Haven: Yale University Press Rajaram, Anand, Tuan Minh Le, Nataliya Biletska, and Jim Brumby. 2010. ―A Diagnostic Framework for Assessing Public Investment Management‖. Policy Research Working Paper: The World Bank Velasco, Andrés. 2000. Debts and Deficits with Fragmented Fiscal Policymaking. Journal of Public Economics 76: 105-25. Wehner, Joachim. 2006. ―Assessing the Power of the Purse: An Index of Legislative Budget Institutions,‖ Political Studies, 54, December Weingast, Barry R., Kenneth A. Shepsle, and Christopher Johnson. 1981. The Political Economy of Benefits and Costs: A Neoclassical Approach to Distributive Politics. Journal of Political Economy 89: 642-64. World Bank. 2010. Mongolia: Improving Public Investment Planning and Budgeting. Washington DC: World Bank World Bank. 2009a. Southern Mongolia Infrastructure Strategy. Washington DC: The World Bank World Bank. 1998. Public Expenditure Management Handbook. Washington DC: The World Bank 22