WPS4379 Policy ReseaRch WoRking PaPeR 4379 Designing Economic Instruments for the Environment in a Decentralized Fiscal System James Alm H. Spencer Banzhaf The World Bank Finance Economics and Urban Department Urban Division & Environment Department Office of the Director October 2007 Policy ReseaRch WoRking PaPeR 4379 Abstract When external effects are important, markets will be assignment of both expenditure and tax responsibilities inefficient, and economists have considered several among levels of government. The standard analysis broad classes of economic instruments to correct therefore focuses mainly upon an aggregate (or national) these inefficiencies. However, the standard economic perspective, it typically ignores the possibility that analysis has tended to neglect important distinctions the externality may be created and addressed by local and interactions between the geographic scope of governments, and it does not consider the implications of pollutants, the enforcement authority of various levels of decentralization for the design of economic instruments government, and the fiscal responsibilities of the levels targeted at environmental problems. This paper examines of government. For example, externalities generated the implications of decentralization for the design in a particular local area may be confined to the local of corrective policies; that is, how does one design area or may spill over to other jurisdictions. Also, local economic instruments in a decentralized fiscal system in governments may be well informed about how best which externalities exist at the local level and in which to regulate or enforce pollution control within their subnational governments have the power to provide jurisdiction, but they may not consider the effects of their local public services, as well as to choose tax instruments actions on other jurisdictions. Finally, the existence of that can both finance these expenditures and correct the locally-generated waste emissions affects the appropriate market failures of externalities? This paper--a joint product of the Urban Division, Finance Economics and Urban Department and the Office of the Director, Environment Department--is part of a larger effort in the Sustainable Development Vice-Presidency to examine policy options for managing urban environment issues. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at prcjra@langate.gsu.edu. 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 DESIGNING ECONOMIC INSTRUMENTS FOR THE ENVIRONMENT IN A DECENTRALIZED FISCAL SYSTEM James Alm and H. Spencer Banzhaf Andrew Young School of Policy Studies Georgia State University Introduction In many settings, markets can be a highly effective way to organize an economy, efficiently making the most of economic resources. In such a system, agents compare private benefits and private costs, ignoring such "external costs" of their actions as pollution. However, when these external effects are important, markets will not be efficient. To correct such deficiencies, economists have considered three broad classes of economic instruments (EIs) (Cointrau and Hornig, 2003). First, if pollution is measurable, setting a "Pigouvian tax" on pollution equal to the marginal external cost of environmental damage would provide the correct incentive to private agents. Second, a "cap" on pollution may be set, with trades or offsets allowed in such a way that the total amount of pollution does not exceed the cap and with the specific agents who do the polluting determined by the market. Third, if pollution is not measurable, "command and control" regulations may be imposed on agents' behaviors, for example by requiring that a factory be required to use a particular pollution abatement technology. There is now a vast literature comparing and contrasting these approaches to controlling pollution. However, the standard analysis in this literature has tended to take the region, and the government, as a given. Consequently, it has tended to neglect important distinctions between the geographic scope of different pollutants, the enforcement authority of various levels of government, the fiscal authority of various levels of government, and the interactions among these factors. For example: 1. Externalities generated in a particular local, or subnational, area may be confined to the local area or may spill over to other jurisdictions. 2 2. Local governments may be better informed about how best to regulate or enforce pollution control within their jurisdiction, but they may not consider the effects of their actions on other jurisdictions. 3. The existence of locally-generated waste emissions affects the appropriate assignment of both expenditure and tax responsibilities among levels of government. The Pigouvian corrective tax (or other economic instruments) may be chosen by the local government in various ways, either making its decision alone and ignoring the actions of other local governments, or making its decision while recognizing that its actions may elicit responses by other local governments as they respond in a strategic manner. The standard analysis therefore focuses mainly upon an aggregate, or national, perspective and typically ignores the possibility that the externality may be created by, and addressed by, local governments. Put differently, the standard analysis does not consider fully, if at all, the implications of decentralization for the design of economic instruments targeted at environmental problems. This paper examines the implications of decentralization for the design of economic instruments. That is, how does one design EIs in a decentralized fiscal system in which externalities exist at the local level and in which subnational governments have the power both to provide local public goods and services and to choose tax instruments that can finance these expenditures and that can correct the market failures of externalities? 1. Externalities and Market Failure: The Standard Approach The standard analysis of an externality starts with welfare economics and the conditions for "Pareto efficiency", defined as a state in which no one can be made better off without adversely affecting someone else (Tresch, 2002). A perfectly competitive market will produce a level of output where private benefits and costs are equal; more 3 precisely, a competitive market will operate where marginal private benefits (MPB) to consumers equal marginal private costs (MPC) to producers. Under some conditions, notably the absence of external effects in either production or consumption, the market will also achieve Pareto efficiency, producing a level of output at which marginal social benefits (MSB) equal marginal social costs (MSC). However, in the presence of external effects like pollution, private benefits and costs diverge from social benefits and costs. In particular, when the production of a product also generates harmful emissions, there will be external, or spillover, costs from production; that is, there will be marginal external costs (MEC) from production of the good, so that MSC, defined as the sum of private and external costs, will exceed MPC by the amount of the MEC. Because a competitive market will produce where MPB = MPC, the presence of pollution now implies that MPB (assumed equal to MSB) equals MPC, but that MSB < MSC MPC + MEC. A competitive market will no longer achieve Pareto efficiency, producing both too much of the product that generates the pollution and doing so in a way that is too "dirty". A comparable analysis focuses more directly on the emissions that characterize the pollution-generating activity, and is useful for subsequent discussion of decentralization. Suppose that the utility of a (representative) consumer depends upon his or her consumption both of a private good and of the level of pollution. Pollution is generated by emissions, or waste discharges, that result from the production of the private good, and production of the private good depends positively upon the use of traditional inputs (e.g., capital and labor), negatively upon the level of pollution, and positively upon the quantity of emissions. Emissions are therefore treated as a factor of production like 4 capital and labor, so that production of the private good can be increased by the use of more capital or more labor, but also by the discharge of more waste because more waste discharges do not require the use of other inputs to limit their use. In this framework, it is straightforward to demonstrate that a competitive market will generate a level of emissions where the marginal private benefits of emissions are driven to zero, because emissions are effectively a "free" input whose use does not require the payment of any factor payments comparable to, say, wages paid for labor. However, the market outcome will lead to an inefficiently large amount of emissions, because private firms that generate the emissions will not consider the external costs of emissions on consumers (who are hurt by more pollution) or on other firms (whose production is also negatively affected by more pollution). However, as noted earlier, there are several corrective policies that a government can pursue (Sterner, 2003). The simplest corrective policy is a so-called Pigouvian tax, equal to the marginal external costs of the pollution at the efficient level of output. The imposition of such a tax brings private costs into line with social costs and thereby achieves the efficient level of production of the good. Polluters who have the lowest cost of abatement will reduce pollution the most, and polluters who have the highest cost will do less abating. In addition to such static efficiency, Pigouvian taxes have the dynamic property of creating an incentive (namely, a reduction in tax payments) for entrepreneurs to develop more efficient ways to reduce pollution. Although they are simple in theory, implementing such taxes in practice raises challenges related to the measurement of the marginal external costs and the monitoring of the pollution levels subject to taxation. 5 Closely related to a tax on pollution is a subsidy on pollution abatement or on cleanup. In this EI, firms face the same marginal incentive as a tax, but the incentive is in the form of a "carrot" (e.g., a subsidy) rather than a "stick" (e.g., a tax). This has the advantage of making the instrument more politically palatable, but it also has the disadvantage of encouraging more activity in a dirty industry, which is counterproductive. Paying for ex post cleanup can be thought of as a subsidy; for example, municipal solid waste collection only occurs because government provides a payment for somebody to do it. Unlike a tax on waste emissions, such cleanup does not reduce the amount of waste generated. Nevertheless, if polluters cannot be identified or if taxes can be avoided by, say, "midnight dumping," there may be no alternative. An alternative EI is a cap-and-trade system, in which a limit is set on the total quantity of pollution allowed. This aggregate cap is then allocated to individual polluters through a permitting process, and may be auctioned or may be distributed without any additional charge. The holders of the pollution permits may in turn trade them to another entity for a mutually-agreed upon price. Although seemingly different on the surface, such a system is really quite similar to a tax: just as under a Pigouvian tax they must pay a tax for each unit of pollution, in the cap-and-trade system polluters must purchase a permit (or must forego the opportunity of selling one). Thus, a cap-and-trade system is formally equivalent to a Pigouvian tax: if a cap allowing pollution level Q leads to permits trading at a price P, then a tax set at P will lead to a level of pollution Q. However, the cap-and-trade system, if properly implemented, guarantees the level of pollution (Q), but not the cost (P); the tax system guarantees that the marginal cost of 6 cleaning up pollution will be P, but does not guarantee that the level of reduction will be Q. The foregoing EIs all work through economic incentives. There are also other types of corrective policies, including ordinary regulation, frequently called "command and control" (CAC). Depending on the nature of the regulation, CAC policies may often be the easiest to administer. For this reason, they are by far the most common approach to pollution control. However, they do not have the same efficiency properties as EIs. For example, they typically impose a one-size-fits-all mandate on all polluters, making them especially burdensome when there is a great deal of variability in the costs for different polluters to clean up. They also typically do not provide incentives for polluters to find better ways to reduce pollution. As summarized by Cointreau and Hornig (2003), these EI can be broadly classified into three categories: revenue-generating instruments, or those EI that produce revenue for governments (e.g., charges, taxes, reductions in subsidies, or auctioned permits), revenue-providing instruments, or those EI that allow producers and service providers to receive income from governments (e.g., fiscal incentives, development rights, or charge/tax reductions), and non-revenue instruments, such as deposit-refund schemes and grandfathered permits. Traditional CAC regulation is also revenue-neutral. 2. Economic Instruments in Context: Legal, Institutional, and Cultural Considerations This standard analysis of economic instruments for pollution control is an elegant application of economic theory, and it provides powerful insights into the advantages of EIs over traditional command-and-control instruments. However, like all basic models, it 7 carries certain explicit and implicit assumptions that must be carefully identified and evaluated before actually implementing any EI, and especially when implementing in a decentralized fiscal system. Adopting an instrument that is cost-effective only under particular circumstances may be quite ineffective if those circumstances do not hold. There are, we believe, five key requirements for EI mechanisms to be effective. First, participants (i.e., individuals or firms whose polluting behavior is to be modified) must be familiar with market prices or, in the context of tradable permits, with market transactions. This requirement is probably the easiest to meet, especially when applied to firms. However, households in some cultures may be more familiar with bartering and haggling, and may be less familiar with a posted price. Even if they are familiar with posted prices in some contexts, if their behavior in, say, water use is guided more by tradition or habit, rather than by constant re-optimization, they may not be responsive to taxes or fees set at politically acceptable levels; even so, the standard theory can accommodate this point by simply describing it as a case of an extremely inelastic demand. Second, in the case of firms, another key requirement for EI mechanisms to be effective is that decision makers must maximize profits. (At a minimum, profit must be one important objective.) Even in highly developed countries such as the U.S. and those in Europe, this point cannot be taken for granted. So-called "principle-agent" problems may arise when actual decision makers do not have the interests of the firm's owners. A plant manager, for example, may care more about making his work-day run smoothly than maximizing the profits of shareholders whom he has never met. In developing countries, especially in communist or former-communist countries, managers may have 8 an administrative mindset in which they focus on meeting output goals rather than profit goals. The potential for this problem is compounded by the presence of "moral hazard", in which counterproductive actions are taken by agents in response to incentives created by government policies. For example, even if managers are keenly watching the bottom line, they may also be shrewdly aware of the fact that costs incurred from paying pollution taxes or fees, or from purchasing pollution permits, can be recovered from governments or from parent companies. Such "soft budgets" (Kornai, 1979) may show up in the form of subsidies, lowered taxes, or more favorably administered prices. By the same token, subsidized pollution abatement may trigger reductions in these side- payments. Western experience shows that such results can be quite subtle. For example, consider the case of U.S. electric utilities. This industry operates under a cap-and-trade system for sulfur dioxide (SO2) pollution, in which the industry receives a certain number of free permits to cover a limited quantity of emissions but which can be bought or sold from one firm to another at prevailing market prices. In a competitive industry, the increased marginal cost of pollution will tend to increased electricity prices, which provides an incentive for consumers to conserve electricity and an incentive also for firms to operate more cleanly. However, the electric utility industry has traditionally operated as a regulated monopoly, charging an administered price for electricity that reflects a fixed mark-up over average costs. Despite a wave of de-regulation in the 1990s, much of the industry continues to operate under this arrangement. Parry (2005) shows that, in this context, increased marginal costs can be offset by the decline in average costs due to the lump-sum transfer of the emissions quota. In such a case, 9 although electricity will be more cleanly produced on a per-unit basis, prices will not increase, and there will be no signal to consumers to conserve energy. In developing countries and, again, especially in communist or former-communist countries, such effects may be quite important. Söderholm (2001) emphasizes the role that administered prices, centrally granted investments, and output targets still hold in Russia, all of which undermine the effectiveness of pollution charges. Bell (2005) argues that the same issues are relevant for China, where much of the industry continues to be directed by the state. A third requirement for EI mechanisms to function properly is the monitoring of pollution. Obviously, if emissions fees are to be collected, or if tradable permits are to be required to be held to cover emissions, those emissions must be observed in a cost- effective manner. For this reason, non-point sources of pollution (such as agricultural run-off, or automobile tailpipe emissions) are rarely if ever covered by EI mechanisms. Rather, even developed countries that embrace EI mechanisms in other contexts use command-and-control regulation such as best-management practices for agriculture and catalytic converters for cars. They may also tax proxies for pollution such as gasoline. Municipal solid waste provides another intriguing example. Although some cities have experimented with charges by weight or volume, such payments can be easily avoided with "midnight dumping". Where EI mechanisms are used, non-revenue generating mechanisms such as a deposit-refund system, or revenue-providing instruments such as a subsidy to recycling, are more common. Here, actors have the incentive to prove their compliance (so as to collect their refund/subsidy). 10 EI mechanisms are best used where they can be observed at a manageable number of discrete points (e.g., filling stations, smoke stacks, or water pipes) and where those points cannot be easily bypassed via "leaks" in the system. Even where most likely to be successful, the continuous monitoring systems that must be installed to monitor and record emissions can be expensive and sophisticated equipment. The expense of such systems is sometimes given as a reason that EI mechanisms are not feasible even for large point sources in poor countries. However, it can also be seen as a concrete opportunity for donors to assist pollution-control efforts. Fourth, enforcement of any violations must be consistent and sufficient to discourage cheating. If polluters, even if detected, know that they will not pay a penalty, then they will not have the needed incentive to reduce pollution. Weak laws, corrupt prosecutors or courts, and a lack or inability to bring legal action on behalf of the public all have the potential to undermine the enforcement of environmental provisions (Bell, 2005). In the case of revenue-generating instruments, such enforcement might be expected to be similar to the enforcement of tax laws. So long as pollution fees are covered by the same branch of law, performance with public finances may indicate the success of future performance with EI mechanisms. There is in fact a large literature on "optimal enforcement" schemes (Kwerel, 1977; Segerson, 1988; Kritikos, 2004). Fifth and finally, it is important for any environmental policy instrument to be accepted by the participants involved. Not all instruments are familiar or acceptable in all cultures. It took many years of discussion before a cap-and-trade system could be tried in the U.S., and more years of experience with early programs before it became widely accepted. These systems still are not used widely in Europe. On the other hand, 11 pollution taxes, used widely in Europe, are used infrequently in the U.S. One can speculate about the reasons: Americans are more willing to commoditize anything and more opposed to taxes, Europeans are more interested in making polluters pay, Americans are more focused on achieving a specific quantity target, and so on (Harrington et al., 2004). Whatever the reason, the important point is that not all instruments will go over equally well in all cultures, even if the economic conditions are the same. We have focused here on requirements for EIs to be fully effective. However, we do not want to leave the impression that, wherever any of these requirements are not fully met, EIs will be a complete failure, or that CAC or other non-revenue and regulatory instruments will be preferable. Each requirement is one of degree, and must be assessed qualitatively. Moreover, many of these considerations are important for CAC instruments as well. For example, a traditional CAC regulation requiring the installation of some specific pollution-reducing equipment too will fail if violations of the regulation are not detected and enforced. Policies must be evaluated according to these criteria on a case-by-case basis, and conclusions will no doubt differ for different societies and different pollution problems. 3. Some Experience with Economic Instruments Before implementing any sophisticated market strategies for reducing pollution, it is essential to examine the experience we have already had with EIs, most of which has occurred in developed countries. Such experience is extensive, and covers an array of instruments for all types of pollutants. In fact, multiple instruments are often used for a 12 single pollution problem in a single country. A recent set of six pairs of case studies, examining the treatment of a particular type of pollution in the U.S. and Europe, found in 11 of 12 cases that a mixture of several tools was used, including both CAC and EIs (Harrington et al., 2004). The U.S. acid rain program, which instituted tradable permits for SO2 emissions on electric utilities, provides an example. All plants, new and old, fall under the cap on total emissions, perhaps the most important EI precedent; Stavins (1998) has called this program "the grand experiment". At the same time, new plants must meet "new source performance standards", which require meeting a maximum rate of emissions per unit electricity, a rate usually set according to "the best available control technology". These performance standards were a legacy of older regulatory approaches kept because of distrust of the EI instrument, and can now be justified as a way to guarantee that trading patterns in pollution permits do not create "hot spots" in a particular area. Unfortunately, they also provide an incentive to keep older, dirtier power plants (which by law are not required to meet the standard) in operation longer. To offset this perverse effect, a process known as "new source review" examines major investments at older plants to determine whether they effectively made the plant a "new source" for regulatory purposes. What can we learn from such an experience? We draw three key lessons; see Harrington et al. (2004) for more details and for additional lessons. First, as textbook theory predicts, EIs are more efficient than CAC regulations. For example, in the U.S. acid rain program, permit prices have traded substantially lower than predicted, suggesting that cost-saving methods of reducing pollution emerged in response to the economic incentives. Carlson et al. (2000) estimate that the cost savings from trading are 13 on the order of $800 million per year, or 43 percent from a uniform emissions rate. Ellerman et al. (2000) similarly estimate the savings to be about 55 percent. In contrast to this success, Sweden's outright ban of trichloroethylene, a toxic solvent used in metal degreasing, met with failure and was repealed. Protests came from a small number of firms experiencing high costs (Sterner, 2004). As noted previously, such cost heterogeneity is a prime example where EIs are preferable to one-size-fits-all regulations. Second, EIs can be more efficient dynamically than CAC instruments. As the U.S. phased out lead in gasoline in the 1980s, it allowed trading to aid small, struggling refineries. It also allowed emissions "banking", in which emissions could be reduced by a greater amount in the present to allow emissions above a (tighter) cap at a later date. Newell and Rogers (2004) estimate the banking alone save more than $225 million, with the total program saving hundreds of millions more. Even without banking, EIs can provide incentives to identify new ways to reduce pollution more cheaply. Such dynamic effects are most likely to be important for highly technical problems such as gasoline refining (where substitutes were found for lead) and air pollution reductions. In the latter case, although there have been few new technologies invented, experimentation with processes (e.g., fuel blending) has led to important discoveries that have reduced costs. This consideration may be less important for situations that require basic actions, or actions that are likely to be part of any cost-effective solution for the foreseeable future. Examples might include basic sanitation or buffers between livestock and water sources. Third, administrative, monitoring, and enforcement costs can differ on a case-by- case basis. As suggested by the Swedish trichloroethylene experience, cost heterogeneity can cause administrative problems for CAC regulations. In contrast, pollution fees and 14 taxes or permits allow more pollution at higher marginal-cost facilities. Many EIs have been implemented in Europe without any reference to firm-level costs (Harrington et al. 2004). As one would expect, monitoring is more costly and demanding for EIs. Again, however, some aspects of the problem can be subtle. For example, in the U.S. lead phase-out, pollution standards were set in terms of lead content (i.e., per unit gasoline). As it turns out, monitoring lead was fairly straightforward because it could be checked against sales figures from suppliers. Gasoline volume (the denominator) was harder to monitor, so a straight cap on lead use rather than on lead content would have been easier to enforce (Newell and Rogers, 2004). Enforcement of EIs has generally been cut and dried. One advantage of these instruments is that they are generally defined objectively, in terms of physical outputs. In contrast, some CAC regulations, such as the use of a "best available control technology", are quite subjective, and require substantial executive rule-making, which in turn in turn may be challenged in court. The U.S. experience with such regulations has involved substantial legal proceedings. In many cases, non-governmental organizations (NGOs) have stepped in to require compliance to fill the void left by executive agencies (Bell, 2005). However, the Europeans, who in general seem less litigious, do not seem to have had these difficulties (Harrington et al., 2004). 4. Implications for the Choice of Pollution-control Mechanisms in Developing Countries As emphasized previously, evaluation of economic instruments must be made on a case-by-case basis. However, the above principles and experiences provide several 15 rules of thumb that provide a starting point for thinking about pollution controls in developing countries. First, in cases where monitoring and enforcement is weak, more objectively measured standards and behaviors should be targeted. Emissions standards, taxes and fees, or CAC regulations requiring very specific actions may all be preferable to standards open to more interpretation. If requirements such as a "best available control technology" or an "adequate margin of safety" create confusion, years of administrative review, and years more of legal challenges even in an open country such as the U.S., they may face even more difficulties in developing countries. Emissions standards or emissions fees, while in some ways quite sophisticated, are also quite objective in their measurement of a physical unit. Here, donors can help by providing or financing expensive monitoring equipment and assisting with their use. In other cases, where enforcement and the rule of law are strong but resources are not available for continuous emissions monitors, trading based on proxies, or modeled emissions, may be a feasible alternative. Montero (2004, 2005) reports success with such a program in Santiago, Chile, where tradable permits are required to be held, not for actual pollution but for a formula based on the technologies employed at each facility. Revenue-providing instruments may also be useful. Far from having an incentive to cheat, in these cases polluters have an incentive to prove their compliance in order to collect a payment. Subsidies thus overcome a monitoring problem, but this comes at an efficiency cost. While they provide the correct marginal incentive for polluters to produce their product in a cleaner way, they subsidize the polluting activity, so that more firms may enter the polluting industry. On net, subsidies thus may not be effective 16 (Baumol and Oates, 1988), especially where firms can easily enter the industry and where abatement costs represent a large fraction of profits. However, they may be quite effective if used in combination with another EI. For example, a deposit-refund is essentially a two-part instrument: a subsidy that encourages recycling combined with an offsetting tax (or the deposit) that eliminates the accompanying perverse incentive to purchase more of the raw material. In the same way, a tax on dirty cars and/or gasoline usage can be used in combination with a subsidy on pollution control equipment to mimic a pollution tax (Fullerton and West, 2002). Second, where the profit motive is weak, rules may be preferable to those mechanisms that rely more upon economic incentives. Emission standards and technology-based standards are consistent with the administrative mindset of firms operating in such contexts, or firms that might not respond to fees or subsidies. As noted previously, such dynamics can arise for firms operating in the context of soft budget constraints, as in the transition countries. Consider another example provided by China. With assistance from the Asian Development Bank, China has experimented with an SO2 pollution trading system in its Taiyuan province (Morgenstern et al., 2005). However, so far the experiment has been a disappointment, with few trades (Predd, 2005). Table 1 summarizes these recommendations. Where the rule of law is strong, polluting activities are observable, and profit motives are strong, the full power of EIs can be unleashed to achieve static and dynamic cost-savings. Where the rule of law is weaker, polluting activities are hard to observe, but profit motives remain strong, subsidies (perhaps in combination with other instruments) may be a better approach. Where the rule of law is strong but the profit motive is weak, traditional CAC regulation 17 may be the best option. Finally, where the rule of law and profit motives are weak, environmental improvements will be particularly challenging. Table 1. Guidelines for the Choice of Economic Instruments Profit Motive Weak Strong Weak; ? Subsidies; Rule of Law; Low Two-part Instruments Observability of Activity Strong; CAC Pollution Fees; High Permits As a final recommendation, we note the importance of guaranteeing that the chosen instrument is accepted by all relevant parties. It took some time for permits to be accepted in the U.S., but this acceptance did not come easily. The provisions allowing trading were scrutinized and tested by skeptics over several years in the 1980s; industry was required to install continuous emissions monitors; permits were assigned serial numbers so they could be tracked and not easily counterfeited; and transactions were recorded. Only after these steps were taken was a consensus reached (Bell, 2005). The lesson we draw from this experience is that, if outsiders are beginning to plan an environmental policy by thinking about the best instrument, then they have already missed a step. The first step, not to be missed, is to think about the right process for choosing an instrument, one that includes local stakeholders. 5. Decentralization: The Assignment of Fiscal Responsibilities 18 The issue of the appropriate assignment of fiscal functions among the levels of government ­ or the optimal amount of centralization or decentralization ­ has long been debated by economists. The Assignment of Expenditure Responsibilities The standard normative discussion of this question follows Musgrave (1959) by dividing the broad functions of government into three branches: the stabilization, distribution, and allocation roles. If the economic criteria for assignment are efficiency and equity, then it is typically concluded that the stabilization function (e.g., smoothing business cycles, reducing inflation and unemployment, encouraging economic growth) should largely be performed by the central government because the mobility of resources makes it unlikely that an effective stabilization policy can be pursued by a lower level of government; local governments also have limited powers to borrow or to print money. Redistribution also properly belongs to the central government. Subnational attempts to redistribute income are likely to be thwarted by the mobility of high-income individuals and of capital, and the attempts to redistribute income will create distortions and inefficiencies in geographic location; the unequal and possibly inadequate fiscal capacities of local governments also make centralization desirable on equity grounds. However, the allocation function, or the decision to provide local government services, should often be performed by local levels of government. These governments can adapt service levels more closely to the preferences of their citizens, thereby making available to individuals a wider range of fiscal choices than could be provided by uniform central government provision. This is the well-known "Subsidiarity Principle", also 19 sometimes referred to as the "Decentralization Theorem" (Oates, 1972, 1993, 1999). An exception may be cases where the service has wide-spread spillovers, such as a public good whose benefits transcend localities. National defense is a classic example here. These considerations suggest a "best practice" assignment of expenditure responsibilities across the different levels of government. Activities such as national defense, monetary policy, and income redistribution are appropriately assigned to the central government; activities like police and fire protection, trash collection, and local roads should be assigned to local governments. Of course, the actual practice on assignment of expenditure responsibilities differs somewhat from these "best practice" assignments. Nevertheless, despite substantial variation in expenditure assignments, the broad principles of assignment are generally upheld in most countries. However, the argument for decentralization is weakened if the costs of local provision are higher due to the smaller scale of local government operations. Importantly, the argument for decentralization is also weakened if there are substantial spillovers from local government expenditures. In this regard, consider "emissions control" or "pollution cleanup" as a government expenditure responsibility. Which level of government should be assigned its responsibility? Put differently, which level of government should be assigned the responsibility for "environmental quality"? Environmental quality is clearly a good that has externality ­ and public good ­ aspects. However, there are different "types" of environmental quality, and the answer to the expenditure assignment question depends on the precise form that environmental quality actually takes. This is discussed in more detail below. 20 Decentralization: The Assignment of Tax Responsibilities Once expenditure responsibilities have been assigned, tax instruments must also be assigned among the levels of government to provide adequate financing for required expenditures. Although there is much diversity in the fiscal structures of national and local governments, several general "best practices" have emerged that provide a useful point of departure: · Only the central government should impose progressive income taxes. Due to the potential mobility of factors, local government attempts to redistribute income by progressive income taxes will lead to the out-migration of mobile, higher-income individuals, thereby leaving immobile, lower-income to bear the burden of the taxes. Income taxes are also thought to be effective countercyclical instruments, and macroeconomic goals are best pursued by national government policies. · The central government should impose taxes on those tax bases that are distributed unequally across jurisdictions, and use the revenues from these taxes to equalize fiscal capacities across these areas. · Local governments should rely predominately upon user charges and taxes on immobile tax bases; in particular, user charges should be used to finance goods that provide measurable benefits to identifiable individuals within a single jurisdiction, and taxes should be used to finance local services for which it is difficult to identify individual beneficiaries and to measure individual costs and benefits. The assignment of taxes should also meet the test of administrative feasibility. · Local governments should avoid taxes on mobile tax bases, especially capital. As with progressive income taxes, the potential mobility of capital or other mobile factors of production will lead to out-migration if these factors are taxed at higher-than-average tax rates. By the same token, attempts to induce in-migration of mobile factors can lead to a so-called "race-to-the-bottom", as local governments compete with each to attract these factors by extending tax breaks and other fiscal incentives. · Local governments should be assigned adequate sources of revenues consistent with their expenditure responsibilities. Local governments should have discretion over the rate of some taxes to promote accountability of local officials and to establish a link between services demanded and the cost of service provision. 21 Locally assigned taxes should exhibit adequate revenue elasticity so that collections can grow with the demand of services over time. · Intergovernmental transfers should be used to finance those services that generate spillovers to nearby jurisdictions, since strictly local finance will lead to inefficient provision. Suppose, for example, that there are two local jurisdictions each of which provides an impure public good (e.g., pollution abatement) whose benefits spill over to the other jurisdiction. It can be shown that each locality should receive a subsidy (e.g., a "conditional", "matching", and "open-ended" grant) on its public good whose magnitude is equal to the marginal benefit of the externality, along the same lines as the earlier Pigouvian tax/subsidy (Oates, 1972; Tresch, 2002). In practice, few countries rigidly follow these guidelines, although the broad pattern of tax assignment is often largely consistent with these prescriptions. Indeed, around the world there are essentially two basic models of revenue assignment that attempt to satisfy these principles. In what might be called the Western or Anglo-Saxon model of "fiscally strong local governments" (e.g., the United States, Canada, Australia), local governments independently legislate and administer their own taxes, an approach that obviously gives local governments significant fiscal autonomy and adequacy. However, this model is probably not appropriate for many countries. Instead, in many other countries the model is one of "fiscally weak local governments" that do not generate much revenues from their own sources, that do not independently legislate and administer their own taxes, but that are often allowed to add a local tax onto the back of some existing central government tax. This approach is being increasingly used as part of decentralization reforms around the world (Bahl and Linn, 1992). It is useful to discuss in more detail the major types of taxes that are used by many local governments, since this discussion relates directly to the design of environmental taxes in a decentralized system. 22 Although there is much diversity in country experiences, the property tax is often a common and important tax for municipal governments, especially those in the Western or Anglo-Saxon tradition. The property tax is in many ways an attractive revenue source. If measured properly, its base should increase with urban growth. Because property can be assessed by physical inspection, the tax is difficult to evade. There is much evidence that the tax has at least a proportional and often a progressive effect upon the distribution of income. The tax is unlikely to create serious distortions in land markets, and may in some circumstances actually improve the efficiency of resource use. Finally, it is sometimes argued that the property tax is most appropriately administered at the local government level because officials there have a better motivation to collect the tax and because the tax can be viewed in part as payment for local services. However, there are also major difficulties with the property tax. The revenue potential of the property tax is seldom realized, due largely to significant administrative problems in identifying properties, valuing them, adjusting valuation over time, collecting revenues, and enforcing penalties. A particular problem is valuation. Property transactions do not occur at regular intervals, which makes it necessary to impose the tax on some estimate of each property's value, and also makes it necessary to use some method to adjust this value over time for changes in prices. Unfortunately, there are few good procedures for such valuation. Experience also demonstrates that it is especially difficult to generate major amounts of additional new revenues from the property tax via short term reforms. The property tax is also often rated by individuals in polls as among the least popular of all taxes. It is highly visible, since the tax is typically paid directly by individual taxpayers rather than by their employer (as with source-withholding of 23 individual income taxes) or by a firm (as with firm payment of sales and excise taxes). Administration of the tax is often highly arbitrary and idiosyncratic, especially in the procedures used to determine the value of properties; when the procedures used to generate this assessed value are performed incompetently, even corruptly, individuals rightly perceive the tax as unfair. The tax base is typically distributed across local governments in very uneven ways, thereby contributing to extreme fiscal disparities across jurisdictions (or "horizontal imbalances"). Perhaps as a result, the property tax is often seen by individuals as a regressive tax, one in which greater burdens are imposed on lower- than on higher-income households. All of these issues are well-known, but this recognition has done little to improve the administration of the tax, even in wealthy countries. Many local governments impose taxes on automobile ownership and use, such as an annual license tax, a registration fee, a transfer tax, a parking fee, tolls, and, at times, a fuel tax (although most countries reserve fuel taxes for central government use). Automotive taxes are a useful source of revenues for local governments. Because car ownership is concentrated in upper income classes, automotive taxes are likely to increase the progressivity of local government finances. Revenues are likely to grow steadily with urban growth. The taxes can be administered at relatively low cost. They can be used for general financing, but they can also be earmarked to finance road construction and maintenance and to decrease congestion and pollution in urban areas. However, with a few exceptions, these taxes are a significantly underused source of revenue. 24 There are several indirect taxes that are potential revenue sources for local governments. Local governments often impose a range of specific excise taxes, sometimes called sumptuary taxes, on commodities like beer, liquor, and tobacco. These taxes generate substantial revenues, they are easy to collect, and they may well discourage consumption of harmful or "immoral" commodities (or "sin taxes"). However, such taxes are also unlikely to grow much over time, they are highly regressive, they may not discourage consumption if consumers are unresponsive to price changes, and their use is clouded by the possibility of individuals buying commodities (or smuggling them from) outside the boundaries of the taxing jurisdiction. Furthermore, collection of excise taxes tends to be concentrated at borders or factory locations, so that they are often unevenly distributed across local governments. It is possible that local governments could obtain some revenues from, say, a central government sales tax, by adding a surtax onto the central government rate, by sharing a specified percentage of the national government collections, or by having a separate retail sales taxes. Some other indirect taxes are not suitable candidates for local government use. These include import duties, manufacturing sales taxes, value added taxes, and other broad-based sales taxes. These taxes are typically reserved for use by higher levels of government. The administration of a value added tax by local governments is particularly complicated and should not be seriously considered, given the usual methods of calculation. Value added taxes are also not suitable for revenue sharing because debiting and crediting of the tax tends to be uneven across regions. Local governments use numerous miscellaneous taxes and fees, which in some cases may be quite important in their finances. For example, many governments impose 25 taxes on various forms of entertainment (e. g., restaurants, hotels, movies, and gambling). Municipal governments also employ a wide variety of "nuisance" taxes. These include stamp duties, license fees, and various minor taxes on advertisements, construction activities, nonmotorized vehicles, and the like. These taxes do not rate highly in revenue performance, administrative ease, and efficiency and distributional effects; they also are seldom an important source of revenues. However, they are likely to continue to be used. Few local governments impose an income tax, at least one administered entirely by the local government itself. An income tax is difficult to administer at the local level, and the central government is seldom willing to share the income tax base with local governments. Given the potential mobility of individuals and firms, the imposition of a local income tax would likely generate significant distortions in resource use. Also, given the predominant use of income taxes for stabilization ends, these taxes should be reserved primarily for central government use. However, it is not uncommon for local governments to impose a surtax, or an additional local government tax, on a national government income tax. The use of such a "piggyback income tax" is a common practice in Scandinavian and central European countries. There are a number of reasons for caution in the use of a local government piggyback income tax. A local government income surtax could generate distortions in resource use, as individuals move to avoid paying the tax and as cities "compete" with one another by changing the tax rate. A local piggyback income tax could also complicate the use of income taxes for stabilization purposes. The fiscal capacity needed to impose a productive local income surtax will differ greatly across localities. The central government may be unwilling to share an income tax base with local 26 governments, especially if (as should happen) the central government has to lower its tax rates to allow localities to piggyback without an overall increase in the tax burden. Still, there are some clear advantages to local surtaxes. The central government administers the tax, thereby avoiding unnecessary duplication of administrative efforts. The central government also retains the authority to define the tax base, which reduces locational distortions from mobile factors and which also reduces interferences of local governments in national stabilization policies, even if these coordination problems are not eliminated. Importantly, local governments are given some discretion in choosing tax rates, within some lower and upper bounds, and this choice enhances their ability to make effective fiscal decisions. Some of these same goals can also be achieved by tax sharing among governments, although tax sharing does not typically give local governments any real authority in the selection of local tax rates and therefore does not promote accountability and efficiency in local expenditures. Surcharges have been increasingly recommended as part of decentralization efforts elsewhere, especially in transition countries where it is necessary to find some fast and sustainable way to give cities a significant fiscal capacity. Of course, tax systems are designed to achieve multiple objectives. An obvious purpose is to raise the revenues necessary to finance government expenditures (sometimes termed "adequacy"), and also to ensure that the growth in revenues is adequate to meet expenditure requirements ("elasticity"). Another is to distribute the burden of taxation in a way that meets with a society's notions of fairness; such "equity" is typically defined in terms of "ability to pay", such that those with equal ability should pay equal taxes ("horizontal equity") and those with greater ability should pay greater 27 taxes ("vertical equity"). Taxes can also be used to influence behavior of those who pay them; in choosing taxes, a common goal is to minimize the interference of taxes in the economic decisions of individuals and firms. Taxes should be simple, both to administer and to comply with because a complicated tax system wastes the resources of tax administrators and taxpayers. The appropriate design of taxes requires balancing tradeoffs among these various goals. However, the implications of locally generated externalities for the appropriate assignment ­ and design ­ of tax instruments have not been fully considered or analyzed. 6. Economic Instruments in a Decentralized Fiscal System The basic analysis of environmental externalities does not take into account the level of government addressing the problem, nor does it take into account either the interactions among jurisdictions at the same level of government or the interactions across levels of a federalist system. By the same token, the standard analysis of the division of fiscal responsibilities and inter-government transfers does not account for the environmental consequences of those relationships. In this section, we draw on both literatures to consider the appropriate role for various levels of government when internalizing externalities, in different settings. What is missing in the standard analysis of externalities? As noted at the beginning, the standard analysis of externalities does not consider that the externality may be generated in a local, or subnational, area, with effects that may be confined to the local area or that may spill over to other jurisdictions. However, the 28 implications of this depend upon the precise form that waste emissions and environmental quality actually takes as a locally produced good. There are at least three different types of polluting activities (Oates, 2002). In a first type, the overall level of environmental quality for the nation as a whole depends upon the aggregate level of waste emissions from all local areas. The actual level of environmental quality in a particular local area may vary across localities, depending on, say, local weather conditions. Even so, however, it is the total amount of emissions from all localities that determines the exact level of environmental quality in any local area. Examples include global climate change and the depletion of the ozone layer, where indeed aggregate worldwide emissions determine the environmental damage at a given location. A second type of pollutant is one in which the amount of waste emissions in a given locality i depends only on the waste emissions in that locality; that is, waste emissions in locality i impose costs on residents and business in locality i, but there are no spillovers from the pollutants beyond the locality itself. An example of this second type of pollutant is the effects of local waste emissions on water quality in the locality, with no spillovers to other jurisdictions (i.e., where the water basin lies entirely within the jurisdiction). A second example is the collection and disposal of municipal solid waste. In a third type of polluting activity, local waste emissions in locality i have harmful effects in that locality, and these waste emissions spill over to at least some immediately surrounding jurisdictions, without necessarily creating an "aggregate" externality; similarly, pollutants from some other neighboring localities ji spillover into jurisdiction i to create a level of environmental quality in i that depends upon emissions 29 from i and from other surrounding jurisdictions. In practice, this third type of waste emissions, which lies between the two previous polar cases, seems likely to be the most common (e.g., acid rain, shared water resources). What is missing in the expenditure assignment analysis? If "emissions control" (or environmental quality) is considered a government expenditure responsibility, then which level of government should be assigned its responsibility? The answer to this question depends upon several considerations that are not fully considered in the standard analysis. A first consideration is which of the three types of polluting activities is being considered. In the first case discussed earlier, where the overall level of environmental quality for the nation as a whole depends upon the aggregate level of waste emissions from all local areas, it is the national government that should be assigned the responsibility for the level of environmental quality, since overall environmental quality depends upon the emissions from all jurisdictions; that is, the provision of overall environmental quality has the characteristic of a pure public good at the national level, and only the central government can (or will) consider the full national effects of waste emissions. However, in the second case, one might argue that it is the local government in each jurisdiction that should be assigned the responsibility for determining the efficient level of environmental quality in its jurisdiction. Because local waste emissions do not spill over to any other localities, the local government should be able to determine the efficient level of the purely local public good, environmental quality. Uniform national 30 standards seem likely to be inefficient because they will not allow local governments to adjust environmental quality either to the demands of local citizens or to local conditions that affect the cost of or optimal approach to pollution control. The third case is the most difficult, where local waste emissions in locality i have harmful effects in that locality and these waste emissions spill over to some (though not all) immediately surrounding jurisdictions. It is unlikely that each locality will consider fully the effects of its waste emissions on surrounding areas, and so it seems unlikely that the efficient level of waste emissions (and of environmental quality) will be achieved in each locality. However, the precise form of market intervention is unclear. A second consideration is that, as noted previously, to work effectively EIs must be backed up by governmental monitoring and enforcement. Central or local governments may differ in these abilities. Depending on how enforcement activities are funded (e.g., out of pollution penalties versus out of general revenues), whichever level of government has more resources may have an advantage. Likewise, whichever level of government has the legal authority to enforce violations will also have an advantage. Finally, levels of governments more free from corruption will be better able to enforce EIs. This may be the central government because it may be less beholden to any particular figure, or it may be local governments because they may be more accountable to local populations. Clearly, the factors discussed earlier will have differing consequences, on a case-by-case basis, for which level of government is most appropriate to operate a given EI, or for which instrument is most appropriate for a given level of government. 31 A third consideration is that in all three types of pollution, but especially in the first and third cases, there is some chance that a local government will choose inefficiently "small" levels of environmental quality, in an attempt to attract mobile capital (e.g., businesses) from other jurisdictions. This is the expenditure equivalent of the so-called tax-side "race-to-the bottom", in which local governments compete with one another to attract businesses, now by lowering environmental standards rather than by lowering taxes. More generally, it has frequently been demonstrated that competition between local governments can lead to inefficient outcomes, largely because local governments attempt to compete for mobile capital by offering fiscal inducements (e.g., lower taxes, lower waste emission standards) to businesses in a race-to-the-bottom. Perhaps surprisingly, however, it can also be demonstrated that, even in the presence of fiscal competition between local governments, local government fiscal choices can achieve efficiency. The crucial difference between these strikingly different conclusions depends largely upon the types of tax instruments that are available to local governments, the composition of the electorate, and upon assumptions made about the presence (or the absence) of strategic local government behavior (Oates and Schwab, 1988). In any event, the possible existence of a race-to-the-bottom in environmental quality is a factor in favor of a more centralized assignment of emissions control (Oates, 2002; Dalmazzone, 2007). The effects of this kind of jurisdictional competition have been tested empirically for at least two cases of the third type of pollutant, in which effects spill over jurisdictions: air quality and transboundary rivers. In U.S. environmental policy, the federal government typically mandates standards of environmental quality that must be 32 met, but allows some degree of discretion on the part of individual states to implement and enforce these standards. In the case of air quality, states were given substantially more discretion (and less assistance from the federal government) in the 1980s under Reagan's "new federalism". List and Gerking (2000) study the effect of this devolution of authority, and find that it had little effect on ­ and, if anything actually increased ­ expenditures on pollution abatement and actual air quality. In the case of water quality, Sigman (2005) similarly finds little over-all effect of more state discretion. However, she does find a small effect of state discretion on downstream states or on rivers that form a shared border. In other words, where the inter- jurisdictional spillovers are strongest, states that have the discretion to do so "free ride." For example, in a worldwide analysis Sigman (2002) finds worse pollution on international rivers than on domestic rivers, with the exception of EU countries where cooperation might be expected to be greater. What is missing in the tax assignment analysis? The analysis of Pigouvian corrective taxation (or other policies) typically focuses upon a single national choice of the corrective tax, and often ignores the possibility that the waste emissions come in different amounts and severities from many local governments; that is, the optimal design of local government corrective taxation (especially in the three possible cases of waste emissions discussed earlier) is an unanswered issue. Indeed, local tax systems in most all countries were originally designed for a world in which production and consumption were primarily of tangible goods, in which 33 the sale and consumption of these goods generally occurred in the same location, and in which the factors of production used to make the goods were for the most part immobile. In such a world, taxation was a fairly straightforward exercise. Sales and excise taxes could be imposed on the tangible goods that were consumed, by the government in the jurisdiction in which consumption (or production) occurred. Similarly, income and property taxes could be imposed on factors where they lived and worked without fear that taxes would drive the factors elsewhere. In making these tax decisions, a government in one jurisdiction had no need to consider how its actions would affect the governments in other jurisdictions because tax bases were largely immobile. There is little doubt that, in principle, decentralization and other associated trends (especially competition among local governments and "globalization", defined loosely as increased factor mobility across jurisdictions) changes things, and changes these decisions dramatically. First, with decentralization tax bases are significantly more mobile across local governments. For example, businesses have more flexibility in choosing where to locate because communication and transportation costs have been slashed. Further, some forms of production activity require little in the way of traditional capital and labor, so that physical location becomes less important. Labor, especially skilled labor, also becomes more mobile in this environment. Likewise, financial capital is able to flow quickly across local (indeed, state and national) boundaries. Clearly, if factors of production can move easily from one location to another, then the ability of a local government to tax these factors is greatly diminished. A government that raises its tax rates above those of other jurisdictions risks losing its tax 34 base to these areas. Particularly in the case of income from capital, there is much speculation that taxation will become increasingly problematic (Mintz, 1992). In fact, there is some empirical evidence (even if not conclusive) that factors are responding to these types of tax considerations (Grubert, 1998; Hines, 1999). Increased mobility is not limited to factors of production. Consumers are also able to plan their consumption according to tax considerations, and consumption does not necessarily occur in the jurisdiction in which a taxpayer resides. A jurisdiction that attempts to tax, say, gasoline more heavily than surrounding areas will find that consumers will purchase elsewhere. Similarly, individuals even in developing countries can now purchase many types of products over the internet and thereby avoid paying some (or even all) sales taxes. Additionally, there has been increased consumption of services and intangible goods (e.g., computer and information services) that are much more difficult to tax than tangible goods. The once-tight link between the location of sales and the location of consumption is now quite loose. Second, and relatedly, the measurement, identification, and assignment of tax bases are now much more difficult. Consider a typical multijurisdictional business. The product that the firm makes may be designed in one or more jurisdictions; the firm may use inputs purchased in multiple jurisdictions; the product may be produced in several places and assembled in a still different location; and the final good may be sold in multiple locations. Because the business operates in multiple jurisdictions, the firm has considerable leeway to manipulate prices to minimize its tax liabilities. This latter problem is well known, but its severity has increased with the enormous expansion in the number of firms operating in multiple jurisdictions. 35 Likewise, consider an individual whose income comes from multiple sources. A global income tax requires that income from these sources be aggregated. However, it is easy for an individual to hide, say, interest income from multiple areas. In the absence of information sharing across governments, the ability of a local government to identify incomes from other jurisdictions is quite limited. Consider finally a consumer who can purchase goods and services in several different ways: from traditional local merchants or from company websites. In the former case, identification, measurement, and assignment of the tax base are straightforward. In the latter case, they are not. Application of, say, sales taxes in this new environment poses considerable problems for governments. How will local governments respond to these various pressures in their tax choices? Most importantly, as emphasized throughout, the ability of any government to choose its tax policies independently of those in other jurisdictions is greatly curtailed. In the presence of mobile tax bases, a single government's choice of tax policies will have effects beyond its own borders and will be affected by the actions of other jurisdictions. Accordingly, the analysis of tax choices by local governments must recognize that a local government will consider that its actions may elicit responses by other local governments as they respond in a strategic manner. These strategic interactions will have a number of effects. The overall level of tax rates seems likely to decline. In particular, if tax bases can move easily from one jurisdiction to another, then they will flow from high-tax to low-tax areas. Owners of capital, skilled labor, and consumers will become increasingly sensitive to tax differentials in their locational decisions. As a consequence, it is 36 commonly argued that governments will face increasing pressures to compete with one another by reducing tax rates or by offering special tax incentives, in order to attract and to retain the various tax bases. For example, when a government reduces its tax rates on capital income, it thereby attracts capital flows from other jurisdictions, and in doing so the government benefits its own jurisdiction. However, the government's action also imposes costs on the jurisdictions that lose factors of production, and it risks generating similar tax-cutting responses from those governments. With tax competition, there could well be a "race-to-the-bottom", in which overall tax collections decline precipitously as local governments compete to attract or to retain their tax bases. To date, however, the evidence here is mixed and inconclusive (Wilson, 1999). The composition of local taxes could also change as a result of increased difficulty in taxing mobile tax bases. The overall tax burden from income taxes on mobile tax bases like capital and skilled labor will likely decline across local governments; tax rates on these factors should also flatten and converge. In contrast, taxes on immobile bases - unskilled labor, physical capital, and property - should increase. Of special note for the use of EIs, charges and fees for specific services should rise in importance because these tax bases are largely immobile. Local governments seem likely to turn more frequently to environmental or "green taxes", as well as to "sin taxes on alcohol and cigarettes and to lotteries, in attempts to replace lost revenues from mobile bases. These compositional changes imply that local tax systems will likely become more regressive than at present. If taxes on capital and skilled labor decline, and if fees and charges, sin taxes, income taxes on unskilled labor, and lotteries all increase, and if 37 marginal income tax rates flatten, then local governments will find it quite difficult to maintain any progressivity in their tax systems. Together with an expected decline in overall revenues, the ability of local governments to redistribute income to lower income individuals will likely diminish. The form of local sales taxes is also likely to change. Local (and other subnational) governments may well decide that a destination-based consumption tax that is collected by the federal government and distributed to them would be preferable to further erosion in their sales tax collections. Alternatively, they may agree among themselves to apply a uniform local sales tax. They may even radically reform the sales tax by moving toward a consumption-based, uniform-rate, destination-principle sales tax, as advocated by McLure (1997) and Fox and Murray (1997) among others. These latter changes suggest more broadly that local governments may attempt greater harmonization (or at least coordination) of their tax systems, in an attempt to reduce the negative (fiscal) externalities that one government's decisions impose upon other governments as well as to harmonize and coordinate environmental policies. Such harmonization implies that there should be some convergence in tax rates across local governments, and also in the definitions of tax bases. With harmonization, local autonomy in tax policy will obviously diminish (Tanzi, 1991). Central governments may effectively induce such harmonization through a system of intergovernmental transfers. If local governments cannot or will not provide adequate environmental protection, say because of competition for mobile capital or because they do not adequately account for inter-jurisdictional environmental spillovers, central governments can increase the level of protection via matching grants (Oates 1999). 38 Similarly, Silva and Caplan (1997) and Nagase and Silva (2000) show that, by making redistribution conditional on local welfare (which in turn may be a function of all jurisdictions' emissions), central governments can induce local governments to take trans-boundary externalities into account. This is a powerful insight, because it suggests that central governments have tools at their disposal that they can employ without knowing anything about abatement costs, the optimal way to abate in a local setting, or even the extent of transboundary externalities. So long as local governments know their externalities, the central government need only observe outcomes. Whether all these changes are good or bad is obviously difficult to determine. Even so, it should be noted that these trends both limit and expand the choices that local governments can make. With greater factor and tax base mobility, local governments have more power to influence the locational decisions of firms, workers, and consumers. Those governments that succeed in these choices will be the ones that are better able to match taxes with expenditures, that are better able to give taxpayers the services ­ including environmental protection ­ that individuals wish for the taxes they pay. Previous research has focused mainly on the negative fiscal externalities of tax competition. It is only recently that the positive effects of tax and, especially, of expenditure competition have begun to be considered in analytical models of local government behavior (Wilson, 1999). It is these positive effects that deserve more emphasis. To illustrate, consider a world in which all factors of production are completely mobile, there are no transportation or communication costs, and there is a single national market for all goods and services. It might seem that no government at any level would 39 be able to impose taxes in such a hypothetical world because any taxes would lead to the immediate outflow of the tax base from the jurisdiction. Put differently, it might appear that the "vanishing taxpayer" would lead inexorably to the virtual disappearance of government. Of course, these assumptions are extreme: complete mobility does not now, and will never fully, exist. Even so, this view is surely wrong. Individuals value the goods and services that local governments provide, and they are willing to pay for them. As originally argued by Tiebout (1956), individuals will "vote with their feet" by moving to those jurisdictions in which governments provide services that residents value. Indeed, local governments will be encouraged, even required, to make their communities as attractive as possible: by providing uncongested roads, a clean environment, pleasant parks, quality schools, safe neighborhoods, and the like, all with a tax burden that individuals deem responsible and appropriate. (If individuals value redistribution, as many certainly do, then even programs for the poor would survive, albeit at smaller levels than currently.) In sum, then, even in an increasingly decentralized (and integrated) economy, local governments will still exist, they will still impose taxes, and they will still make expenditures. There is no question that these decisions will be circumscribed by the possibility of a "vanishing taxpayer". However, the existence of such a taxpayer also creates opportunities, by giving local governments the potential to influence these locational decisions. Local governments whose prior performance has been poor will have little credibility in making policy decisions; the response to those governments with sound institutions will be quite different. There will therefore be pressures on all local 40 governments to establish these institutions. Those governments that succeed in these choices will be the ones that are better able to match taxes with expenditures, or are better able to give taxpayers the services they desire for the taxes they wish to pay. Conclusions There is by now a large literature on the fiscal constraints and opportunities in a federal system. However, there has been less attention paid to the choice of environmental instruments in such a system. We believe that there is much that can be learned by bringing these literatures together. We conclude by drawing several lessons from this synthesis. Our first perspective relates to the choice of which level of government should implement environmental instruments when such a choice is open. Out second perspective relates to the choice of which instrument might be most fitting at different levels of government. If a fully functioning federal system offers some choice as to the level of government that should take action to reduce environmental pollution, our first and most important principle is the geographic scope of the externality. If the effects of waste fall within the same jurisdiction as the source, then local governments are probably best situated to address the externality. If the waste has significant transboundary effects, then national governments are better positioned to address it. To this first principle, we add several mitigating factors. First, environmental policies must of course be enforced, so levels of government that have stronger police and judicial powers and that are freer from corruption are preferable. Second, even if the effects of pollution are local, if local sources of pollution are constituents of a national 41 industry, then there may be efficiencies in imposing national rules. Third, and again even if the effects of pollution are local, local governments may be unable or unwilling to tackle the pollution problem because of fiscal competition. If it is a question of taxing the pollution or polluting industry, then local governments may fear the loss of jobs in that industry. Further, if it is a question of raising taxes on other sources in order to raise funds for (say) cleanup or for a revenue-providing instrument, then local governments may fear the loss of that alternative tax base. In these cases, central governments may prove more effective or may reduce these problems through revenue transfers. Again however, it is important to recognize the importance of a "race-to-the-bottom", which remains an open question. Consider now the alternative perspective on the problem: the choice of environmental instruments when the level of government is already determined. One lesson here is perhaps an obvious one but one that must nevertheless be stated clearly. Namely, the instrument must be consistent with the fiscal authority of the government. In particular, a revenue-generating instrument must be tied to a treasury with the authority to levy and manage those taxes. Our second principle is that the instrument should be consistent with the fiscal needs of the level of government. If the jurisdiction is in need of revenues, a revenue- providing instrument will naturally be more appealing that a revenue-generating instrument. Third, and similarly, the instrument should depend on the mobility of the polluting industry. Because polluters are naturally more free to move across local borders than to move across national borders, central government may have more ability 42 to employ revenue-generating instruments on polluting industries, whereas local governments may face stronger political pressures to keep the industry. The local government may therefore prefer to use a revenue-neutral or even revenue-providing instrument. At the margin, these can provide the same incentive to polluters to reduce their pollution but can impose a lower total burden. However, local governments must also think about the mobility of the polluting industry relative to other tax bases. The funds for a revenue-providing instrument must come from somewhere. If the other sources of revenue are even more mobile than the dirty sources, it would not make sense to raise taxes on them further. Indeed it would make more sense to employ a revenue- raising instrument for pollution and reduce those other taxes. Fourth, EIs are likely to have their strongest advantage in the context of national policies. The reason for this is that much of their efficiency arises from the fact that they allow more abatement to be borne by polluters who have lower marginal abatement costs, and heterogeneity in these costs is likely to be greater the wider the geographic scope. As Adam Smith himself noted, the gains from trade are limited by the extent of the market. At the extreme extent of a pollution market, the arrival of global environmental problems such as climate change have posed well-known challenges for environmental policy-makers. However, as the world's population continues to urbanize at the same time as governments are trending toward more decentralized authority, more local concerns must not be forgotten. Such pollution problems as local sanitation and water quality continue to be of tremendous human importance and raise policy challenges of their own. Given this range of environmental problems facing policy-makers, no single policy prescription could possibly be adequate for all problems. While the "right" policy 43 choice will always be highly contextual, the literatures on environmental instruments and fiscal federalism provide important insights into the problem. References and Relevant Readings Bahl, Roy W. and Johannes F. Linn (1992). Urban Public Finance in Developing Countries. Oxford University Press and The World Bank. Barthold, Tim (1994). "Issues in the Design of Environmental Excise Taxes". The Journal of Economic Perspectives 8 (1): 133-152. Baumol, William J. (1972). "On Taxation and the Control of Externalities". The American Economic Review 62 (3): 307-322. Baumol, William J. and Wallace E. Oates (1988). The Theory of Environmental Policy. New York, NY: Cambridge University Press. Bell, Ruth Greenspan (2005). "Culture ­ and History ­ Count: Choosing Environmental Tools to Fit Available Institutions and Experience". Indiana Law Review 38: 637-69. Bergstrom, Theodore (1976). "Regulation of Externalities". Journal of Public Economics 5 (1): 131-138. Bird, Richard M. (1999). "Rethinking Tax Assignment: The Need for Better Subnational Taxes". IMF Working Paper, Washington, D.C. Bovenberg, A. Lans and Ruud A. de Mooij (1994). "Environmental Levies and Distortionary Taxation". The American Economic Review 84 (4):1085-1089. Bovenberg, A. Lans and Lawrence H. Goulder (1996). "Optimal Environmental Taxation in the Presence of Other Taxes: General Equilibrium Analyses". The American Economic Review 86 (4): 985-1000. Bovenberg, A. Lans and Lawrence H. Goulder (2002). "Environmental Taxation and Regulation". In Alan J. Auerbach and Martin Feldstein, Handbook of Public Economics, Volume 3. Elsevier ­ North Holland Publishing. Braden, John B., Henk Folmer, and Thomas Ulen (eds.) (1996). Environmental Policy with Political and Economics Integration: The European Union and the United States. Cheltenham: Edward Elgar Publishing Co. 44 Breton, Albert (1998). Competitive Governments: An Economic Theory of Politics and Public Finance. New York, NY: Cambridge University Press. Carlson, Curtis, Dallas Burtraw, Maureen Cropper, and Karen L. Palmer ( 2000). "Sulfur-Dioxide Control By Electric Utilities: What are the Gains from Trade?" Journal of Political Economy 108 (4): 1292-326. Cointreau, Sandra and Constance Hornig (2003). "The Application of Economic Instruments in Water and Solid Waste Management". Regional Policy Dialogue, Inter-American Development Bank. Washington, D.C. Cornes, Richard and Todd Sandler (1986). The Theory of Externalities, Public Goods, and Club Goods, Cambridge University Press. Cropper, Maureen and Wallace Oates (1992). "Environmental Economics: A Survey", The Journal of Economic Literature 30 (4): 675-740. Dalmazzone, Silvana (2006). "Decentralization and the Environment." In Ehtisham Ahmad and Giorgio Brosio (eds.) Handbook of Fiscal Federalism. Cheltenham, UK: Edward Elgar, 459-77. Dinan, Terry M., Maureen L. Cropper, and Paul R. Portney (1999). "Environmental Federalism: Welfare Losses from Uniform National Drinking Water Standards". In Arvind Panagariya et al. (eds.), Environmental and Public Economics: Essays in Honor of Wallace E. Oates. Cheltenham: Edward Elgar Publishing Co., 13-31. Ellerman, A. Denny, Paul L. Joskow, Richard Schmalensee, Juan-Pablo Montero, and Elizabeth M. Bailey (2000). Markets for Clean Air: The US Acid Rain Program. Cambridge: Cambridge University Press. Engel, Kirsten H. (1997). "State Environmental Standard Setting: Is There a `Race' and Is It `To the Bottom'?" Hastings Law Journal 48: 271-398. Farber, Daniel A. (1997). "Environmental Federalism in a Global Economy". Virginia Law Review 83: 1283-1319. Fullerton, Don, and Sarah E. West (2002). "Can Taxes on Cars and on Gasoline Mimic an Unavailable Tax on Emissions?" Journal of Environmental Economics and Management 43 (2): 135-157. Gordon, Roger (1983). "An Optimal Tax Approach to Fiscal Federalism". Quarterly Journal of Economics 97: 567-586. Gramlich, Edward M. (1987). "Subnational Fiscal Policy". Perspectives on Local Public Finance and Policy 3 (1): 3-27. 45 Hahn, Robert (1989). "Economic Prescriptions for Environmental Problems: How the Patient Followed the Doctors Orders". The Journal of Economic Perspectives 3 (1): 95-114. Hahn, Robert (1990). "Regulatory Constraints on Environmental Markets". Journal of Public Economics 42 (2): 149-177. Harrington, Winston, Richard D. Morgenstern, and Thomas Sterner (eds.) (2004). Choosing Environmental Policy: Comparing Instruments and Outcomes in the United States and Europe. Washington, D.C.: Resources for the Future. Hazilla, M. and R. Kropp (1990). "Social Cost of Environmental Quality Regulations: A General Equilibrium Analysis". The Journal of Political Economy 98 (4): 853- 872. Kornai, János (1979). "Resource-Constrained Versus Demand-Constrained Systems". Econometrica 47 (4): 801-819. Kwerel, Evan (1977). "To Tell the Truth: Imperfect Information and Optimal Pollution Control". Review of Economic Studies 44: 595-601. Kritikos, Alexander S., (2004). "A Penalty System to Enforce Policy Measures under Incomplete Information". International Review of Law and Economics 24 (3): 385-403. List, John A, and Shelby Gerking (2000). "Regulatory Federalism and Environmental Protection in the United States." Journal of Regional Science 40(3): 453-471. McLure, Charles E. Jr. (1994). "The Tax Assignment Problem: Ends, Means, and Constraints". Australian Tax Forum 11: 153-183. McLure, Charles E. Jr. (ed.) (1983). Tax Assignment in Federal Countries. Center for Research on Federal Financial Relations. Montero, Juan-Pablo (2004). "Tradable Permits with Incomplete Monitoring: Evidence from Santiago's Particulates Permits Program". Mimeo, Cambridge, MA: MIT. http://tisiphone.mit.edu/RePEc/mee/wpaper/2004-015.pdf. Montero, Juan-Pablo (2005). "Pollution Markets with Imperfectly Observed Emissions". The Rand Journal of Economics 36 (4): 645-659. Morgenstern, Richard D., Piya Abegunawardena, Robert Anderson, Ruth Greenspan Bell, Alan J. Krupnick, Jeremy Schreifels, Cao Dong, Wang Jinan, Wang Jitian, and Steiner Larsen (2005). "Emissions Trading to Improve Air Quality in an Industrial City in the People's Republic of China". In China's Environment and 46 the Challenge of Sustainable Development, Kristen A. Day (ed). Armonk, NY: M.E. Sharpe. Musgrave, Richard A. (1983). "Who Should Tax, Where, What?". In Charles E. McLure, Jr. (ed.), Tax Assignment in Federal Countries. Center for Research on Federal Financial Relations, 2-19. Nagase, Yoko, and Emilson C. D. Silva (2000). "Optimal Control of Acid Rain in a Federation with Decentralized Leadership and Information." Journal of Environmental Economics and Management 40(2): 164-80. Newberry, David (1990). "Externalities: The Theory of Environmental Policy". In Hughes and Heal (eds.), Public Policy and the Tax System, 106-149. Newell, Richard G, and Kristian Rogers (2004). "Leaded Gasoline in the United States: The Breakthrough of Trading". In Harrington et al. (ed.), 175-191. Oates, Wallace E. (1972). Fiscal Federalism. Harcourt Brace Jovanovich. Oates, Wallace E. (1993). "Federalism and Government Finance". In John M. Quigley and Eugene Smolensky (eds.), Modern Public Finance. Boston, MA: Harvard University Press, 126-151. Oates, Wallace E. (1999). "An Essay on Fiscal Federalism". The Journal of Economic Literature 37 (4): 1120-1149. Oates, Wallace E. (2002). "A Reconsideration of Environmental Federalism". In John A. List and Aart de Zeeuw (eds.), Recent Advances in Environmental Economics. Cheltenham, UK: Edward Elgar, 1-32. Oates, Wallace E. and Robert M. Schwab (1988). "Economic Competition among Jurisdictions: Efficiency-enhancing or Distortion-inducing?" Journal of Public Economics 35 (4): 333-354. Parry, Ian W.H. (2005). "Fiscal Interactions and the Costs of Controlling Pollution from Electricity". Rand Journal of Economics 36: 849-869. Poterba, James (1993). "Global Warming Policy: A Public Finance Perspective". The Journal of Economic Perspectives 7 (1): 47-63. Predd, Prachi Patel (2005). "A Market for Clean Air: China Adopts Emissions Trading to Curb Pollution". http://envirovaluation.org/index.php?cat=202. Revesz, Richard I. (1992). "Rehabilitating Interstate Competition: Rethinking the "Race- to-the-bottom". New York University Law Review 67: 1210-1254. 47 Segerson, Kathleen (1988). "Uncertainty and Incentives for Non-point Pollution Control," Journal of Environmental Economics and Management 15 (1): 87-98. Sigman, Hillary (2002). "International Spillovers and Water Quality in Rivers: Do Countries Free Ride?" American Economic Review 92(4): 1152-9. Sigman, Hillary (2005). "Transboundary Spillovers and Decentralization of Environmental Policies." Journal of Environmental Economics and Management 50(1): 82-101. Silva, Emilson C. D., and Arthur J. Caplan (1997). "Transboundary Pollution Control in Federal Systems." Journal of Environmental Economics and Management 34: 173-86. Söderholm, Patrik (2001). "Environmental Policy in Transition Economies: Will Pollution Charges Work?" The Journal of Environment and Development 10: 365-390. Stavins, Robert N. (1998). "What Can We Learn from the Grand Policy Experiment? Lessons from SO2 Allowance Trading". Journal of Economic Perspectives 12 (3): 69-88. Sterner, Thomas (2003). Policy Instruments for Environmental and Natural Resource Management. Washington, DC: Resources for the Future. Sterner, Thomas. (2004). "Trichloroethylene in Europe: Ban versus Tax". In Harrington et al. (ed.), 206-221. Tiebout, Charles (1956). "A Pure Theory of Local Expenditures". The Journal of Political Economy 64 (3): 416-424. Tresch, Richard W. (2002). Public Finance ­ A Normative Theory. Academic Press. Wellisch, Dietmar (2000). The Theory of Public Finance in a Federal State. Cambridge University Press. Wilson, John D. (1999). "Theories of Tax Competition". National Tax Journal 52 (3): 269-304. Zodrow, George and Peter Mieszkowski (1989). "Pigou, Tiebout, Property Taxation, and the Underprovision of Local Public Goods". Journal of Urban Economics 19 (3): 356-370. ED4Admin C:\Documents and Settings\wb14621\My Documents\green economics\WB-DecentralizationEnvironment&Taxes-WB DEC version.doc 07/26/2007 5:14:00 PM 48