Abstract
The global community faces a number of critical challenges ranging from climate change to crossborder health risks to natural-resource scarcities. Many of these so-called global commons problems carry grave risks to economic growth in the developing world and to the livelihoods and welfare of their people. Climate change is the classic example. Despite the risks involved, donor governments have funded programs addressing global challenges such as climate change at far lower levels than traditional programs of country-based development assistance. The prospects for dealing with such global challenges will depend at least in part on new collective financing mechanisms. In this paper, we examine four categories of existing resource-mobilization options, including (1) transportation levies; (2) currency and financial transaction taxes; (3) capitalization of IMF Special Drawing Rights (SDRs); and (4) the sale, mobilization, or capitalization of IMF gold. In the end, we recommend that willing governments utilize a modest portion of their existing SDR allocations to capitalize a third-party financing entity. This entity would offer bonds on international capital markets backed by its SDR reserves. The proceeds would back private investment in climate-mitigation projects in developing countries that might otherwise lack adequate financing.
This approach could mobilize up to $75 billion at little or no budgetary cost for contributing governments. Any limited budgetary costs could be offset by using excess proceeds from recent IMF gold sales. In our view, capitalizing a small portion of existing global assets—SDRs with a small back-up reserve of the income from gold already sold—to finance programs that deal with global public goods and bads makes eminent sense.

I. INTRODUCTION
The global community faces a number of critical common challenges, including climate change, increasing cross-border health risks ranging from pandemic flu to drug resistance, and politically- and economically-destabilizing natural resource scarcities, such as local and regional access to water. Many of these so-called global commons problems carry grave risks to economic growth in the developing world and to the livelihoods and welfare of their people. In welfare terms, those problems pose the greatest risk to the poorest countries and the poorest people.
For decades, traditional donors have committed resources to address global challenges that matter acutely in the developing world. In the early 1970s, the Consultative Group on International Agricultural Research (CGIAR) was created, which played a critical role in the Green Revolution in South Asia. In the 1980s, the Global Environment Facility (GEF) was created. Since then, the GEF has supported programs and projects in the developing world for protection of biodiversity, forests, marine life, and other related issues. More recently, the donor community has created Climate Investment Funds at the World Bank and the other multilateral banks.
However, these initiatives have been funded at far lower levels than traditional country directed development assistance programs. Official aid for country-based programs in 2009 amounted to about $120 billion.1 In that year, our rough estimate of official transfers to support non-country based global programs – such as basic agricultural research, vaccine production and distribution, UN peacekeeping, preserving biodiversity, and reducing greenhouse gas emissions – amounted to less than $12 billion (see table 1 and appendix 1 for additional details). In comparison to traditional development aid, revenue mobilization to finance global public good programs has been limited.2 In contrast to traditional development programs, spending by the traditional donors on these latter programs does not provide the same returns politically or diplomatically as spending on bilateral aid. In this context, the political dynamics raise the same or even more difficult challenges domestically as committing resources to the United Nations and other multilateral institutions.
Despite the apparent difficulty of raising revenue for global public goods (GPGs), the rising profile of climate change and the pressure to achieve some concrete outcome at the UN Conference in Cancún in December 2010 triggered substantial new financial pledges by the developed countries. The developed country signatories agreed to provide $30 billion by the end of 2012 ($10 billion a year) and to establish a Green Climate Fund that would mobilize $100 billion annually by 2020 to assist with climate adaptation and mitigation.
There is a general assumption that much of the $100 billion will somehow be generated through carbon trading (in the form of developed country transfers to developing countries to offset the former group’s emissions above some agreed level). That is evident for example in the November 2010 report of the UN High-level Advisory Group on Climate Change Financing (hereafter UNAGF – in which some estimates of private resources that could be raised assume a market price of $25 per ton of emissions.) However, the reality is that even with global trading of emissions rights, the rich world will still need to raise considerable public resources to meet its announced commitment. This is obviously the case for adaptation, which requires public resources. But even for mitigation, public resources are needed. That is obviously true in the short run, as a global cap and trade system is not in the cards in the next few years. And in the longer run, large transfers through private trading to developing countries (say on the order of $50 billion a year) depend on initial allocations of emissions rights that are politically unlikely7 (as well as an implicit price of at least $25 ton).
Finally, even if the world does manage to create a carbon trading system involving such large transfers, there will still be many additional opportunities to deploy public resources to cover the gap between social and private returns that trading will not cover, if only because of the need to establish reporting and verification in developing countries which is critical to private trading (but can be managed in the case of large investments that are publicly subsidized).
Many would argue that the first-best choice is for each country’s political processes to authorize and appropriate the necessary funds. However, in the fallout of the global economic crisis, donor governments face significant political as well as fiscal constraints to expanding development assistance budgets over the next several years at least. Given this, the prospects for helping developing countries deal with new, shared GPG challenges would seem to rest, at least in part, on new collective financing mechanisms that are specifically designed to address GPGs as distinct from “development”, i.e. goods that directly or indirectly reduce risks or produce benefits important to rich as well as poor countries. For these GPGs, particularly climate mitigation, we take the view that it is time to begin developing political support for new pools of financing that are inherently collective or cooperative at the global level. That support will need to come from key governments as well as from citizens, civil society, legislative bodies, private businesses, and other stakeholders that influence such support, particularly though not only in the rich world.

In Section II of this paper, we examine four options for increased financing of GPGs. By definition, these GPGs have benefits for rich countries as well as poor countries, while at the same time are, as we say in our title, development-pertinent. In other words, they are critically important for shared growth and development in developing countries. We envision in particular programs for climate change mitigation, though new technologies and investments in agriculture, energy, water, and health also clearly would fall under the GPG category.
The four options are: (1) transportation levies; (2) currency and financial transaction taxes; (3) capitalization of IMF Special Drawing Rights (SDRs); and (4) sale, mobilization or capitalization of IMF gold. We provide background on each option, and then briefly assess the technical merits and political feasibility of each. All of these except use of gold are discussed in the UNAGF report; we refer briefly below to findings of that report on each.
In particular, we highlight legislative and other requirements pertaining to the United States – whose participation is critical either to avoid free riding or arbitrage, or in the case of aviation levies, to reach a reasonably robust level of financing. In Section III, we outline a specific proposal for further analysis and advocacy. We focus on the advantages and likely challenges of securing agreement from willing IMF members to more effectively harness modest amounts of their existing SDR assets.
The resulting resources would be utilized to catalyze private investments with high social compared to private returns that lack financing despite their commercial viability (even at current low carbon prices and more so if and when those prices rise), for climate mitigation in developing countries.
Our proposed use of SDRs would not address the need for public resources for climate adaptation financing in developing countries. Other grant-based or highly concessional resources would be required for adaptation as well as other GPG programs that are development-pertinent – including research and development of new technologies in agriculture and health. However, it would help meet a portion of the immediate and longer-run commitments (of $30 billion and $100 billion), in principle easing the pressure on total budget demands for other global programs.
At the end of the paper, we note that specific institutional arrangements for deploying any resources raised through any of the collective or cooperative options have not been addressed. In the case of the United States and possibly many emerging market economies, garnering political support for any of the proposals will require simultaneous agreement on what existing or new institutional arrangements would be used. In this context, governance and management structures of any new international facility are key issues.
II. COLLECTIVE FINANCING: FOUR OPTIONS
The options we explore, as mentioned above, are all “collective or cooperative” at the global level. The aviation levies and capitalization of existing SDRs are “cooperative”. They do not require that all countries agree, though they do require agreement of the large economies if they are to succeed in raising a reasonable amount of revenue – e.g. on the order of $10 billion annually – and cooperation on rates and deployment of the resources. The maritime, currency, and financial taxes (and use of IMF gold) require that all countries agree either due to existing governance structures or to avoid arbitrage by market players.
A. TRANSPORTATION LEVIES
Description: This paper assesses options for maritime and aviation levies coupled with, or independent of, emissions trading schemes (ETS) within each industry.13 These measures could be implemented concurrently with aviation ticket fees or levies, either at a flat rate or ad valorem. For approach descriptions and revenue estimates, we draw extensively from the United Nations High-Level Advisory Group on Climate Change Financing (UNAGF) report published in December 2010 and an IMF working paper focused on international aviation taxes.
Overall, we believe that maritime levies display some promise in terms of revenue mobilization as well as providing positive environmental externalities. Aviation ticket levies have a major advantage of permitting cooperative action instead of requiring universal adoption. However, the revenue mobilization prospects are modest without the participation of the United States and other major countries and would not discourage emissions, unless structured in complicated and controversial ways.
Maritime Fuel Emissions Taxes: Under the UNAGF proposal, a limit would be set on greenhouse gas emissions in the international maritime sector along with a predetermined, fixed price for carbon credits based on international consensus. Maritime carriers would either purchase these credits as emissions offsets (so that as in a cap and trade system the total amount of emissions under the scheme would be fixed) or would simply pay the tax on fuel used. An agreed percentage of revenues from credit sales would be directed towards climate financing. The UN Panel recommends a universal levy as opposed to a differentiated levy given that any differentiation by route or flag would generate inefficiencies as parties would seek to minimize their tax payments. The Panel reports that the International Maritime Organization (IMO) likely would not oppose a well-designed scheme, especially if it is built on existing monitoring and recordkeeping requirements. According to the Panel’s mid-range estimates, a maritime fuel tax could generate between $4 billion and $9 billion annually by 2020. A market-based approach – which would allow for a floating carbon offset credit price –would mobilize a slightly lower level of revenues.
Aviation Fuel Levies: We examine two types of aviation fuel-related levies: (1) excise taxes; and (2) emissions taxes coupled with a carbon trading scheme. Aviation Fuel Excise Tax: Many countries already apply a value-added tax (VAT) on domestic aviation fuel. In practice, the effective taxation rate ranges dramatically across countries. Given a restrictive international legal framework, few countries have taxed aviation fuel used on international flights. Theoretically, signatory countries could re-open existing international conventions and agreements to explicitly permit international aviation fuel taxes. However, this would be time intensive and complex (particularly with bilateral agreements), which makes this highly unlikely in the near- or medium-term. According to the IMF, an aviation fuel excise tax of $0.20 per gallon could yield up to $9.5 billion annually based upon: (1) aviation fuel usage figures in 2003; and (2) global imposition on all domestic and international flights.
Aviation Fuel Emissions Tax: Under an alternative approach, a cap could be placed on greenhouse gas emissions generated through aviation fuel usage. Subsequently, carbon offset credits would be auctioned at an international market-based price. Based on UNAGF analysis, this approach could generate roughly $2 billion annually by 2020. As with the maritime fuel emissions tax, the UN Panel argues that a universally-imposed levy would be more politically acceptable than an approach differentiated by country or carrier. Some industry players may prefer a cap and trade-type approach (i.e., floating carbon offset price) to a fuel excise tax since it allows greater flexibility via permit trading and grandfathering.
Aviation Ticket Taxes: In 2006, several countries – led by France – committed to apply modest levies on airline tickets to finance HIV/AIDS, tuberculosis, and malaria programs in developing countries.25 To date, seven countries have implemented a “solidarity” tax on airline tickets – including: Chile, Côte d'Ivoire, France, Madagascar, Mauritius, Niger, and South Korea.26 The tax is applied to domestic and international flights departing from their respective territories and fixed rates vary across participating countries. In 2009, the solidarity air ticket tax mobilized roughly $170 million – with France accounting for nearly 95 percent of the total. Since its inception, it has generated nearly $700 million after tax collection expenses.28 Tax proceeds are channeled to UNITAID, an international drug purchasing facility housed at the World Health Organization. UNITAID purchases AIDS and other drugs for use in over 90 low-income developing countries. Under a global airline ticket tax scheme, the UNAGF estimates that revenues could total between $0.5 billion and $5 billion annually by 2020.30 Alternatively, the IMF has estimated that revenues could reach roughly $10 billion through an average tax of $6 per passenger.31 If the ticket tax were applied only in Europe, the IMF estimates that revenues could reach up to $2.4 billion annually.
Possible Resource Usage: Globally, there are no inherent limitations on how these international transportation-related levies could be used. Sponsoring governments have wide latitude in determining individually how and where to channel their tax receipts – including for domestic purposes. It is worth noting again that the precedent for channeling aviation ticket and fuel tax proceeds for a non-domestic purpose have been established.
U.S. Legislative Requirements: As with any new tax measure, imposition of new nationwide maritime or aviation taxes would require new congressional legislation. The probability that the U.S. Congress would pass these taxes and earmark a significant portion of revenues for non-domestic usage is very low in the current political environment.
Global Externalities: Maritime and aviation travel contributes to local air pollution through the release of nitrous oxide, carbon monoxide, hydrocarbons, sulphates, and soot aerosols.33 It also contributes to global warming through carbon dioxide emissions– though relatively modestly compared to other sources of carbon and other greenhouse gas emissions. For example, the 1999 Intergovernmental Panel on Climate Change estimated that aviation would account for only about 5 percent of global carbon emissions by 2050 (one to two percentage point increase from mid-2000 levels). While this is a very small share, it would not be insignificant in absolute terms if other emission sources decline. As designed, the emissions-based tax options applied to either the aviation or maritime sectors would not only mobilize financial resources, but also establish incentives for operators to lower greenhouse gas emissions. On the other hand, maritime and aviation levies at higher levels could reduce international trade and tourism.
Approach Strengths: Since the mid-2000s, major donor governments have explored and debated the potential role of aviation taxes. Maritime taxes have not received broad, senior-level attention until more recently. Key strengths or advantages include: (1) Linkage with Global Public Goods: Due to fuel-related emissions, there is a case for using aviation or maritime levy proceeds to finance climate change mitigation and other GPGs. (2) Low Price Elasticity: The maritime and aviation sectors’ relatively low estimated price elasticity of demand indicates that levies could have a marginal dampening effect on overall maritime trade volumes. (3) Tax Administration: Preexisting levies (i.e., security fees, departure fees, etc.) illustrate that aviation and maritime levies are feasible and entail relatively modest administrative expenses. For a universally-implemented ETS, compliance requirements would be relatively simple. (4) Offsets Flexibility: An ETS-type approach for either maritime or aviation fuel would provide operators with flexibility to purchase carbon credits from other sectors. (5) Revenue Mobilization: Maritime and aviation proceeds would present a reliable source of finance. They are best seen as an option unless and until there is a globally agreed tax or cap at which time they would be absorbed into a larger trading or tax system.
Approach Weaknesses: Legal and policy limitations have prevented more widespread imposition of additional levies by developed and developing countries. (1) Legal Restrictions: In the case of aviation excise fuel taxes for international flights, the Chicago Convention and bilateral air service agreements pose a significant obstacle. Some U.S. bilateral agreements might also include legal restrictions on air ticket taxes. (2) Collective Action Challenges: Except in the case of aviation ticket taxes, which exist already on a voluntary basis for some countries, the other transportation levies (with or without trading), would have to be universally agreed. Agreement would also be needed on the proportion of nationally collected revenue to be used for international or global programs as opposed to domestic purposes. (3) Impact on Airline Carriers and Tourism: The financial health of many airline carriers remains weak. High fixed and variable costs (mainly fuel) – coupled with constrained consumer demand – have prevented a full-scale recovery after sector-wide difficulties in the early- and mid-2000s. Depending on their ultimate size, new aviation levies could provide an additional headwind against sector-wide recovery and health. Effects on tourism might also be a problem, particularly for small island economies and other countries highly dependent on that industry. (4) Industry Opposition: Though the maritime industry has signaled the levies could be implemented at low cost, both aviation and maritime players are likely to object to a tax that does not cover other emission-generating industries, such as power generation or road transport. (5) An Earmarked Tax: Utilizing transaction tax proceeds for non-transportation related uses violates the textbook view that earmarked taxes are bad in the first place and worse if the use of revenue raised is unrelated to the sector taxed. (6) Domestic Revenue Collection and Competition with Provision for AIDS and Other Diseases in Poor Countries: In the case of the ticket tax which need not be universal, political competition for resource usage beyond the current arrangement – which assigns the resources to UNITAID – likely would scuttle the tax itself in a country like the United States. The key question is whether“global” uses could be expanded and whether that would affect resources for AIDS medications and other lifesaving drugs.
B. CURRENCY AND FINANCIAL TRANSACTION TAXES
Global development advocates have for decades focused on taxing various types of financial transactions. The most attractive advantage is their high transaction volume and, by extension, the potential to raise large amounts of revenue through a very low tax rate. The “Robin Hood” tax website includes extensive discussion and commentary on the revenue potential of these taxes, while emphasizing their likely progressive nature. In this manner, if the tax was passed on to “consumers”, they mostly would be banks and other financial firms (including those doing proprietary trading) and presumably relatively well-off investors.
In this section, we briefly examine the two most widely discussed variations: (1) financial transactions taxes; and (2) currency transaction taxes (variations of the “Tobin tax”). Among the transaction taxes, we see the most merit in what is called the centrally collected multi-currency tax (at a low rate of 0.005 percent).
We assess the advantages and disadvantages of these taxes assuming a very low rate – one that is unlikely to throw sand in the wheels of international capital flows, which was an initial objective of a Tobin tax (see below), or to check what might be speculative transactions. Its sole objective would be raising revenue.
Currency Transaction Tax (CTT): Originally proposed in 1978, the so-called Tobin Tax was intended to apply an internationally-uniform tax on all spot currency conversions. The tax would be proportional to the size of the transaction and be administered by individual governments. The central objective was to improve national monetary policy effectiveness in a post-fixed exchange rate environment. Tobin argued that increased mobility of private capital could lead to excessive cross-border movements – including those of a speculative nature – which would produce significant economic costs. Although Tobin acknowledged distortional and allocation costs, he argued that they would be small compared to the macroeconomic costs of excessive international capital mobility.
Recent currency-related transaction tax proposals have focused on revenue mobilization as opposed to reducing speculative flows or broader financial and macroeconomic policy issues. Given this, the proposed tax would be set at a very low rate in order to minimize market distortions. According to Schmidt (2008), a tax of 0.005 percent on the four most widely traded currencies (U.S. dollar, Euro, Yen, and Pound) could produce roughly $33 billion annually.43 However, earlier studies suggest lower annual receipts ranging between $17 billion and $24 billion.44 A new study by the Institute for Development Studies estimates annual receipts of roughly $26 billion.
Centrally Collected CTT: A 2009 international task force report includes revenue mobilization estimates for a multi-currency transaction tax collected centrally. The tax would be applied to all transactions – regardless of currency denomination – that are settled through the Continuous Linked Settlement (CLS) system. The CLS system, which is operated by the CLS Bank, is a process by which the largest global financial institutions manage settlement of foreign exchange.48 Currently, the CLS system is available to seventeen foreign currencies, settles an estimated 55 percent of all spot, swap and forward transactions, and has an average daily value of payment instructions of $3.8 trillion.49 The estimates range between $25 and $34 billion annually based on different spreads and price elasticities.
Financial Transaction Tax: Building on a currency transaction measure, a financial transaction tax could also include equities, bonds, and derivatives. This approach would not put currency markets at a relative disadvantage to other asset classes. According to the Austrian Institute of Economic Research, a comprehensive financial transaction tax of 0.01 percent could generate between 0.8 percent and 2.0 percent of global GDP ($410 billion to $1.06 trillion). Derivatives-based revenues account for the vast majority of these tax proceeds. Without them, the related tax revenue could range between $72 billion and $80 billion annually.52 By extension, the Austrian Institute’s methodology would suggest that a financial transaction tax of 0.005 percent would produce between $210 billion and $530 billion annually (roughly $35 billion to $40 billion without derivatives). As noted previously, significant caution should be given to these revenue estimates due to a number of methodological factors.
U.S. Legislative Requirements: As with any new tax measure, imposition of currency and/or more general financial transaction-related taxes would require new congressional legislation. The U.S. Treasury has discouraged discussion of transaction tax measures in international fora.
美国立法要求:对于任何新的税收措施,实施的货币和/或更一般的金融事务相关的税收需要新国会的立法。美国财政部已气馁讨论事务税收措施在国际论坛。
Approach Strengths – the Centrally Collected Multi-Currency Tax: There is a long, exhaustive, and contentious discourse on the relative pros and cons of imposing taxes, on certain types of financial transactions. We focus on the strengths and weaknesses of the centrally collected multi-currency tax, with some notes comparing it to other financial transactions taxes. A few advantages of a modest (0.005 percent) centrally collected currency tax are: (1) Revenue Mobilization Potential: It likely would raise significant sums of money, with little projected impact on the size and volume of centrally-settled transactions because of the offsetting existing advantages of central settlement. (2) Low Administrative Cost: Administrative costs would be low. This also would be true of a financial transaction tax; administrative costs associated with the British stamp tax total roughly 0.21 pence for every pound collected (in contrast to 1.24 pence for the income tax). (3) Collective Collection: By avoiding national collection and disbursal (as would be the case for the bulk of any financial transactions tax), the centrally collected tax overcomes the so-called domestic revenue problem. (4) Equitable Burden Sharing to Finance a Global Public Good: The impact of the tax would be proportional across countries and currencies based on level of engagement in international markets. It would represent a tax associated with the benefits obtained from use of the global commons that a liberal international order provides. In contrast, a financial transactions tax would fall largely on the United States, the United Kingdom, and a handful of other countries.
Approach Weaknesses for a Centrally-Collected Multi-Currency Tax: Weaknesses or challenges include: (1) Cyclical Revenues: Tax receipts would be volatile and pro-cyclical – with greater receipts in economic expansions and reductions in downturns. Of course, this shortcoming could be addressed through expenditure smoothing over time. (2) Collective Action Challenge: Virtually all countries involved in international trade and capital movements would need to agree on the tax. This also is the case without central collection for any currency or financial transactions tax. All countries would need to impose the tax measure; otherwise, there would be arbitrage involving non-participating states. (3) Potential Inconsistency with Existing Treaties: Transaction taxes may be incompatible with the General Agreement on Trade in Services (GATS) as well as the Treaty on the Functioning of the European Union.57 If a new international treaty were needed to override existing agreements, it would imply a long and complex international negotiation. (4) An Earmarked Tax: Utilizing transaction tax proceeds for non-financial uses violates the textbook view that earmarked taxes are bad in the first place and worse if the use of revenue raised is unrelated to the sector taxed.
C. IMF SPECIAL DRAWING RIGHTS
SDR Overview: One potential resource mobilization approach is the direct or indirect utilization of Special Drawing Rights (SDRs). SDRs are an international reserve asset created by the IMF in 1968 to supplement other reserve assets of member countries. They are valued on the basis of a basket of currencies and can be used in a variety of transactions and operations among official holders. Under its Articles of Agreement, the IMF may allocate SDRs to members in proportion to their IMF quotas. Put differently, SDRs are established through a collective global decision-making process (i.e., the IMF Governors) for a collective purpose (i.e., international financial liquidity and stability).
In financial terms, the IMF administers each member country’s SDR account. These members pay an applicable SDR interest rate on their allocations while also receiving interest payments on their actual SDR holdings. This process produces a revenue neutral outcome as long as a member country’s SDR allocation and holdings are the same. If a member’s SDR holdings rise above its allocation (because it has acquired SDRs from other SDR holders), then it earns net interest on its excess SDRs. If a member country holds fewer SDRs than its allocation (having sold or transferred SDRs to other holders), then that country pays interest on its shortfall.
Over time, the IMF has authorized, implemented, and/or agreed to four SDR allocations for its member countries (see figure 5 below). In August 2009, the IMF approved an unprecedented allocation of SDR 161.2 billion (roughly $240 billion) to member countries. This allocation was designed to expand international reserves and liquidity in response to the global economic crisis. Of this, nearly $150 billion was distributed to developed countries. (Also in August 2009, the IMF executed an additional special allocation through an amendment to its Articles of Agreement, which primarily m benefited members that that had not received earlier SDR allocations – such as countries of the former Soviet Union. That increased the total cumulative allocations to SDR 204 billion).

An IMF member country may use SDRs freely, without the requirement of need, to obtain an equivalent amount of currency through a voluntary agreement with another member country. For these SDR transactions, the IMF acts as an intermediary between member countries and those SDR holders that have voluntarily agreed to participate in SDR transactions. By illustration, Uganda could decide to sell some of its existing SDR 144 million holdings to China in exchange for U.S. dollars. In turn, the Bank of Uganda could decide to hold these U.S. dollars as reserves and/or pursue expansionary fiscal policies (i.e., purchase Ugandan treasury bills). However, the Ugandan government would have lower SDR holdings after this transaction and would therefore pay interest on its shortfall.65 The Chinese government would receive an increase in the interest payments on its higher SDR holdings.
As of June 2010, only one member country (Tuvalu) had SDR holdings equivalent to its allocation (see appendix 5 for details). This illustrates that SDRs have been traded regularly over time. In the event of SDR market illiquidity, the IMF can activate a formal designation mechanism. Under this mechanism, members with sufficiently strong external positions are designated by the IMF to buy SDRs with freely usable currencies up to certain amounts from members with weak external positions, but this mechanism has not been used for over three decades.
SDR-Based Financing Approaches – Description: SDR-based financing has been done or discussed in one of two forms: (1) monetizing SDRs, either through SDR on-lending or in freely usable currencies following conversion – several countries have agreed to lend a portion of their SDRs to the Poverty Reduction and Growth Facility, which provides concessional loans to low-income members, such as Haiti; and (2) committing SDRs to support the capitalization of a third-party entity. The latter has been proposed in an IMF staff paper, which we discuss below. Of these two, we focus on the capitalization approach since it could mobilize significant resources at a low cost to traditional donors.
SDR Capitalization Approach: Under this approach, any member country wishing to participate would make an ongoing commitment of a limited portion of their SDR holdings (e.g. 10 percent) to capitalize a third-party entity. (10 percent of SDR holdings of the developed plus the many advanced developing countries would amount to about $15 billion.) This entity would offer bonds on international capital markets backed by its SDR reserves. Assuming that the entity does not incur substantial losses, the SDR capital would never be monetized or formally utilized. As a result, sponsoring governments would not incur SDR interest expenses.
Bredenkamp and Pattillo (2010) proposed that (only) developed countries utilize their portion of their latest general allocation of SDRs (over $100 billion) to provide capital for a new Green Fund.71 The Green Fund could be simply a resource mobilization instrument that would disburse to existing climate funds at the multilateral banks and/or through any new agreed Global Green Fund. The bond proceeds raised against the SDR capital, which would command AAA, would be used to help finance commercially viable (though sometimes at lower returns than so-called dirty alternatives) climate mitigation investments through loans, guarantees, and other financial backing. Bredenkamp and Pattillo assume that the resulting official financing of investments could be leveraged by private sector equity and lending at a ratio of 10:1. The recent UN AGF report and related background documents suggest a more modest leveraging ratio of 3:1 or 4:1).
Possible Resource Usage: In the context of a SDR-backed financing option, there are no explicit restrictions on how IMF member countries use their SDR holdings. Sponsoring governments of a capitalized fund (such as a Green Fund) would have wide latitude in determining how and where to channel their holdings – certainly including financing climate and other global public goods. IMF member countries would of course be setting the precedent of utilizing a portion of their existing SDR holdings in the interest of “climate stability” and of other GPGs; our argument is that these are fundamental to long-term global financial stability.
Cost Implications: Sponsoring governments’ costs ultimately depend upon the specific structure of the SDR-based approach as well as country-specific factors, such as budgetary system regulations. The idea is that the structure would be one in which the SDRs that countries allocate would be viewed as completely unencumbered and thus still treated as each sponsoring country’s reserve assets. If that structure were not adopted fully, then there could be up to three types of costs: (1) fixed budgetary costs related to committing and scoring SDR allocations; (2) variable SDR interest expenses; and (3) foreign reserve requirement costs.
First, IMF member countries typically record SDR holdings as foreign reserves. For some countries, offering these holdings directly or indirectly to a special entity could entail budgetary costs. The Bredenkamp and Pattillo proposal builds in sufficient capital to ensure the SDRs committed by countries would always be available to them and would thus continue to qualify as reserve assets – fulfilling the core global financial stability mandate of individual member countries’ SDR holdings. The ultimate budgetary impact would depend on: (1) how member country governments “score” capital commitments on their budgets; and (2) the timeframe over which the SDRs would be provided or committed.
Second, governments could incur interest expenses were the SDRs “called” by the third party entity in excess of their preexisting SDR holding levels. By illustration, sponsoring governments would have incurred approximately SDR 380,000 in interest expenses between 2000 and 2010 for every SDR 1 million worth of deficit SDR holdings that were provided upfront. The costs would generally be borne by taxpayers.
However, as we propose below, in the highly unusual event of a “call” on sponsoring governments’ SDRs (there has never been a call on IMF or World Bank capital by their creditors), they could plan to cover interest costs through coordinated implementation of other collective financing approaches, such as IMF gold sales. Such an agreement could be structured a number of different ways, such as setting an interest rate expense ceiling for governments or offsetting a specific percentage point volume of expenses (e.g., covering 2 percentage points of the overall SDR rate).
Lastly, some governments conceivably could confront broader foreign reserve adequacy issues. For them, using their SDR holdings could require a proportional increase in other reserve currencies. Overall, SDRs account for only 6 percent of developed plus advanced developing countries’ official reserve holdings (see appendix 7 for details).
So it is unlikely that reserve substitution costs would probably be an issue.
U.S. Legislative Requirements: Utilizing SDRs to directly or indirectly finance GPG programs almost certainly would require congressional amendments to existing legislation. There are several overlapping U.S. legislative acts that govern the allocation and usage of SDR holdings. Under the Special Drawing Rights Act of 1968, the U.S. Congress gave the Executive Branch authority to vote for SDR allocations that meet specific conditions.83 Currently, SDRs allocated by the IMF or acquired by the U.S. Government are considered resources of the Exchange Stabilization Fund (ESF) of the U.S. Treasury.84 As of April 2010, SDR holdings accounted for approximately 55 percent of the ESF’s assets.85 In accordance with the Special Drawing Rights Act, the ESF (through the Secretary of the Treasury) can monetize SDRs by issuing certificates to the Federal Reserve.86 In turn, the ESF may use these resources to finance exchange stabilization operations. Historically, the ESF has provided short-term currency swap agreements or loans to developing countries experiencing a currency or debt crisis. There are also precedents of medium-term loans as well as loan guarantees.
Given the inconsistency with the ESF’s mandate, utilization of SDRs for the purpose of capitalizing a GPGs facility could raise policy concerns for the U.S. Even if only a modest portion of U.S. SDR holdings were committed, and even if the commitment were structured so that it would in principle be totally unencumbered, the U.S. Treasury Department (and U.S. Congress) likely would be concerned about tying up ESF assets that would otherwise provide stability in the event of a financial crisis overseas.
Global Externalities: Spillover benefits depend upon the specific structure and programmatic focus of the SDR-backed financing approach. For example, the Bredenkamp and Pattillo proposal could boost the incentive for establishing a functioning carbon trading market.
Approach Strengths: SDR-backed financing approaches provide a number of strengths in terms of budgetary implications, policy outcomes, and opportunity cost tradeoffs. (1) Resource Maximization: SDRs offer a significant pool of available capital. Currently, IMF member governments’ allocations total approximately SDR 204 billion ($307 billion) – of which, developed countries hold roughly SDR 129 billion ($195 billion). We propose a target total of $15 billion. (2) Existing Asset Base: Given that IMF member countries already hold significant SDR allocations, new resource mobilization (i.e., new allocations) is not required– in contrast to currency and financial transactions taxes, as well as aviation levies. (3) Limited Foreign Reserve Role: On average, SDRs account for roughly 6 percent of developed countries’ official reserve assets, and less of the reserve assets of China, India, Brazil and Russia. If their SDR allocations were re-deployed in their entirety, developed countries’ import cover would only decline, on average, from 5.4 months to 5.1 months. (4) Modest, if any, Interest Costs: IMF member countries would incur interest expenses only if their SDR allocation had to be called and exceeded their preexisting holdings. Currently, the SDR interest rate is roughly 0.30 percent. Between 2000 and 2010, the interest rate averaged 3.55 percent.
Approach Weaknesses and Risks: At the same time, sponsoring governments must consider several weaknesses or challenges related to pursuing SDR-backed financing approaches. (1) Upfront Budgetary Cost: At least in the case of the United States, there could be a decision to score the cost if SDRs were utilized – even if unencumbered. (2) Non-Monetary System Usage Precedent: Some IMF shareholders will have concerns about using SDRs for a purpose not immediately linked to monetary policy or global financial stability. Though scoring does not require appropriations, it does get included in the budget and would add to the announced deficit. (3) Legislative Requirements: For the U.S. government, congressional approval would be required. As noted above, the use of ESF funds for non-economic crisis purposes might raise concerns within the U.S. executive branch – independent of views within the U.S. Congress. (4) Interest Expenses: In the event that SDRs were called, any relevant interest expenses would have to be paid by sponsoring governments until they bought back the SDRs or unless they were paid by other revenue sources such as IMF gold sales. Although, these expenses likely would be modest on an annualized basis. (5) Uncertain Commercial Viability of Funded Programs: The SDR-based capitalization approach is predicated upon recipient programs and investments yielding a financial return sufficient to cover the fund’s cost of capital and administrative expenses. Given this, loans would have to be commercially viable, though they would not need to achieve as high a rate of return as private investors might otherwise demand.

D. IMF GOLD SALES
IMF Gold Overview: As of October 2010, the IMF held 91.5 million ounces (2,847 metric tons) of gold at designated depositories. The IMF’s total gold holdings are valued on its balance sheet at SDR 3.4 billion (about $5.3 billion) on the basis of historical cost. At the current high market price of gold compared to the past, the IMF's holdings amount to approximately $114 billion – or roughly $109 billion more than its balance sheet valuation.95 The IMF has acquired its gold holdings through several types of transactions over time. First, it required that its member countries provide 25 percent of their initial quota subscriptions and subsequent quota increases in the form of gold. Second, the majority of members’ payments of loan interest charges were made in gold prior to 1978. Third, the IMF in the past would facilitate member countries’ access to foreign currencies in exchange for gold (again prior to 1978). Lastly, member countries could use gold to repay the IMF for credit previously extended. Through the latter modality, the IMF acquired 13 million ounces following the Second Amendment of the IMF's Articles of Agreement in April 1978. It is these post-1978 gold acquisitions, whose use is governed by the Second Amendment, that are of interest for our purposes.

According to the 1978 Second Amendment, the IMF has the authority to: (1) sell gold holdings outright on the basis of market prices; (2) accept gold in the discharge of member countries’ loan repayment obligations; and (3) “restitute” gold to members, in particular, it can sell gold at the former official price (SDR 35 per ounce) to those countries that were members as of August 31, 1975 in proportion to their quotas on that date. Each of these transactions requires Executive Board approval by an 85 percent majority. With a voting share exceeding 15 percent, the United States has a de facto veto on those gold operations. The Second Amendment also prohibits utilizing IMF gold for financial operations, such as loans, leases, swaps, or collateral. As a result, any gold-based financing approach requires that IMF gold holdings are first sold outright or “mobilized”.
IMF gold holdings do not affect its normal operational lending capacity. Instead, its lending capacity is driven by two forms of usable resources, including: (1) quota-funded currency holdings of financially-strong economies; and (2) borrowing from its own members and the market under explicit member-approved facilities.
IMF-Based Financing Approaches - Description: IMF gold-based financing proposals could take three forms -(1) outright sale of existing gold holdings; (2) “mobilizing” existing gold holdings through off-market transactions; or (3) amending the Articles of Agreement to allow the IMF to utilize its gold holdings to capitalize a third-party entity. (1) IMF Gold Sales Approach: Since its founding in 1944, the IMF has sold gold on several occasions. These transactions have directly financed operations, among other things, as:
Currency Replenishments (1957-70): The IMF sold gold on several occasions to replenish its holdings of currencies.
U.S. Government Securities Investments (1956-72): Modest IMF gold holdings were sold to the United States to generate revenue to cover operational deficits. The sale proceeds were invested in U.S. government securities. Later, the IMF reacquired the respective gold from the U.S. government.
"Restitution" Sales (1976-80): Following agreement by IMF member countries to reduce the role of gold in the international monetary system, the IMF sold roughly 50 million ounces. Half of these sales occurred through member country restitution procedures. Under these procedures, member countries could purchase gold acquired before the Second Amendment (April 1978) at the former official price of SDR 35 per ounce (see additional details below). Restitution sales require support from an 85 percent majority of the total voting power.
Auction Sales (1976-1980): The remainder of gold sales during this period was executed through market auctions to finance the Trust Fund. This Trust Fund, which later morphed into the Poverty Growth Reduction Trust (PRGT), supported concessional lending to low-income countries.
Most relevant for our purposes, however, is the recent decision to sell gold to help finance non-income generating IMF administrative or operational costs especially ongoing surveillance of membersˇ countries financial and macroeconomic policies.
Surveillance Activities and Low-Income Country Lending Capacity (2009): In September 2009, the IMF Executive Board formally approved the sale of all gold holdings acquired after the 1978 Second Amendment (12.97 million ounces) or roughly one-eighth of its holdings. The original purpose was to reduce IMF dependence on lending income to cover operational expenses.100 Prior to the global economic crisis, lending volumes and associated revenues had declined dramatically, thus creating operating budget shortfalls. Income generated from the gold proceeds (through investments in conservative instruments) is meant to finance a portion of the IMF budget related to its provision of GPGs (i.e., surveillance). Following the onset of the global economic crisis, the IMF Executive Board also agreed that a portion of the income on the endowment from the sale of the gold would be used to subsidize interest rate expenses for concessional lending to low-income countries through 2011.
As noted, the 2009 agreement entails the sale of all gold holdings acquired after the 1978 Second Amendment. These sales were completed in December 2010. Due to the continued run-up of gold spot prices since mid-2009, the IMF’s gold sale proceeds exceeded the forecasted requirements for the surveillance endowment and supplemental low-income country lending. However, the IMF has not yet publicly disclosed the transaction-specific or average sale prices – and therefore, the size of the excess proceeds. Any surplus resources, which likely total several billion dollars, will remain unallocated until the IMF Board of Directors decides how to use them. We propose below that one possible usage would be to cover costs associated with increased financing of climate and other development pertinent GPGs.
（2）Mobilizing Gold Holdings Approach: Alternatively, IMF member countries could “mobilize” gold holdings. In simple terms, this approach serves to monetize the differential between prevailing gold market prices and the accounting valuation placed on the IMF’s holdings (see figure 7 below). At a gold spot price of about $1,250 (compared to an accounting valuation of roughly $50), IMF member countries could generate up to roughly $110 billion for development-pertinent GPGs from the IMF’s remaining gold holdings.105 Importantly, these transactions would not impact global gold prices since no IMF gold actually enters international markets, which could lead to a decline in global prices.

The IMF utilized this approach to provide funding for the Heavily Indebted Poor Countries (HIPC) Initiative in 1999. At that time, the IMF Executive Board authorized off-market transactions with Brazil and Mexico totaling up to 14 million ounces. First, the IMF sold 12.9 million ounces to these countries at the prevailing market price. The resulting profits then were placed into a special HIPC Initiative account. Second, Brazil and Mexico provided the same amount of gold back to the IMF to settle their separate loan obligations with the IMF. As a result, the IMF’s overall gold holdings remained unchanged. However, the IMF’s holdings of currencies were reduced by the amount of the gold profit, which also reduced its lending capacity for non-concessional financing (such as to Mexico or Greece). By extension, this reduces its income as the IMF then has correspondingly larger interest paying liabilities (reserve positions) to member countries. If all existing IMF gold holdings were “mobilized”, then the IMF would assume a liability of up to $110 billion on its balance sheet. Given its interest paying nature, this liability could produce significant financial and operating implications for the IMF.
Under this approach, IMF member countries would relinquish their claims to the pricing differential. Given the financial implications, this could entail budgetary implications for some IMF member countries depending on national budgetary regulations and scorekeeping rules. (2) Capitalization Approach (Article of Agreement Amendment): Lastly, member countries could agree to amend the Articles of Agreement (i.e., Second Amendment) to explicitly permit the IMF to utilize pre-1978 gold holdings to capitalize a third-party entity. As with the SDR capitalization approach, the respective third-party entity would utilize the IMF gold as reserves and raise operating capital by floating bonds on international markets.
This approach also could entail important budgetary implications. Ultimately, the respective treatment would depend on such factors as member countries’ budgetary scorekeeping regulations and the risk profile associated with the third-party entity’s activities.
Possible Resource Usage: In the context of gold sales or mobilization, there are no explicit restrictions on how proceeds may be used.109 As with capitalization of SDRs, IMF member countries have wide latitude in determining how and where to channel any resulting resources – ranging from financing GPGs to targeted country or sector programs. IMF member countries would need to consider the opportunity costs of utilizing these resources for development-pertinent global public goods as opposed to other potential uses, such as additional debt relief for low-income countries.
U.S. Legislative Requirements: Under existing U.S. law, the executive branch can support the sale of IMF gold without congressional action only if the Secretary of the Treasury certifies to Congress that a sale is necessary either for the restitution of gold to IMF member countries; or to provide liquidity enabling the IMF to meet member countries’ claims generally or threats to international financial stability. Otherwise congressional authorization is explicitly required.
Effect on Gold Market Prices: IMF gold sales need to be modest, gradual, and transparent to avoid a major decline in market prices. Even then, some risks of gold commodity market disruption would remain. However, we would expect that IMF member countries with important gold production or reserve holdings (i.e., Canada, Australia, Germany, Russia, South Africa, United States, etc.) would play an important role in shaping any sales and communication strategy, to minimize the potential market impact.
Approach Strengths: IMF gold-based financing approaches provide a number of strengths: (1) Revenue Mobilization Potential: IMF gold-based approaches could yield significant financial resources for development-pertinent GPGs – especially if the proceeds are used to capitalize a third-party entity. In terms of immediate revenue potential, outright sales could produce up to $104 billion while mobilization could yield roughly $100 billion. (2) Use of An Existing Asset Base: Given that the IMF (through member countries) already has significant gold holdings, new resource mobilization is not required (in contrast to financial transactions and aviation levies). Currently, IMF gold is an under-utilized asset that provides no operational benefit in terms of institutional lending capacity. (3) A Source for Grant Funding (Though Only Once): Gold mobilization would generate resources that could be provided as grant financing (e.g. to low-income countries for climate adaptation). In contrast to capitalization, the resulting stock of resources could only be used once.
Approach Weaknesses: Gold sales or mobilization have weaknesses for our purposes. Some of these weaknesses apply to using gold to capitalize a fund as well. These include: (1) Upfront Budgetary Cost: As noted previously, some governments could incur significant upfront budgetary costs related to IMF gold sales, mobilization, or capitalization operations. (2) Legislative Requirements: For the U.S. government, congressional approval would be required – which may present a significant obstacle to concrete action. (3) Potential Gold Market Impact: Anything other than limited and gradual gold sales could cause a decline in gold market prices, reducing the value of member countries’ official reserve holdings. Countries with large or relatively concentrated gold reserve holdings (i.e., France, Germany, Italy, and Netherlands) likely would oppose large sales, so the revenue potential of using IMF gold (other than as callable capital) is limited in practical terms. (4) Impact on IMF Lending Capacity and Income: The “mobilization” of gold by transferring accounting profits from the IMF’s holdings of currencies provided by members’ quota subscriptions would reduce its lending capacity by the same amount, which would have ongoing budgetary costs for the IMF.
III. A SPECIFIC PROPOSAL
By way of conclusion, in this section we consider how to approach these options in a particular case: that of climate change and the likely U.S. financial obligations. The climate change case is apt because it requires a politically visible and specific commitment by the developed countries of $10 billion annually for developing countries between 2010 and 2012; and a rapidly mounting commitment after that (although much less specific in terms of financing sources). These out-year commitments increase to $100 billion annually by 2020. The United States is implicitly committed to provide about 25 percent of these amounts (some may argue that it should provide an even larger share).
Among the four options discussed above, which ones are technically practical and politically feasible? Particularly for the United States, whose participation is indispensable given its size and influence in the global economy? Each approach has its merits and should be considered and debated by policy advocates and civil society groups. In the near term, additional consideration of the maritime fuel tax – with or without emissions trading – likely makes sense and potentially could mobilize up to $5 billion to $10 billion annually while also reducing carbon emissions.
Nonetheless, to ensure a larger pool of resources for climate and other GPGs that matter for development, we believe it makes sense to tap existing global assets in a manner that has minimal direct budget costs for the advanced economies. We specifically emphasize the advantage of some combination of: (1) seeking an agreement among IMF members to more effectively harness limited amounts of existing SDRs to capitalize a Climate Mitigation Fund (using SDRs committed by any members wishing to participate but requiring some minimum number to trigger creation of the fund) for use in developing countries; and (2) utilizing a limited amount of excess resources from the agreed sale of IMF gold to help meet any future SDR-related interest expenses.
SDRs Plus Some Gold: We believe the most attractive option, for the United States and for the global community as well, is the utilization of a very modest portion of SDRs to capitalize a third-party entity. In turn, this entity would provide lending for climate friendly and potentially commercially-viable projects (along the lines of Bredenkamp and Pattillo, 2010). We can imagine advanced country and emerging market members of the IMF designating 10 percent of their cumulative SDR allocations for this purpose (totaling roughly 10 billion SDRs or almost $15 billion). Depending on the third-party entity’s ability to leverage those financial backstop resources (we think a leveraging ratio of 5:1 is reasonable given the experience of the World Bank Climate Investment Funds), the facility could mobilize as much as $75 billion for climate mitigation projects and programs in developing countries.
Conceptually the logic of using a limited amount of SDRs to finance an agreed global program (i.e., Cancún climate financing commitments) is straightforward. SDRs constitute a global asset designed to help stabilize the global economic and financial system in the event of various shocks. While financial stability traditionally has been paramount in the minds of the finance and central bank officials who influence decisions at the IMF, climate change poses a direct and indirect threat to financial and geopolitical stability – particularly given its unpredictable risk profile over time and across countries. Minimizing the resulting uncertainty and risks using SDRs would contribute to global stability. In today’s international system, stability must be broadly defined to include the risks of food price and other commodity price shocks, of unpredictable and sudden large movements of people, of weather-related humanitarian crises and so on.
This proposed usage of SDRs would not require upfront agreement by all IMF member countries. Instead, a coalition of sponsoring governments could press forward initially. However, it would be difficult to imagine its success without the agreement or political backing of the United States – including the U.S. Congress.
Countries deploying some of their SDRs in this fashion could face variable interest charges if the third-party entity ever encountered significant loan losses. However, conservative financial management and loan review policies would minimize this likelihood. If participating countries did confront SDR interest charges, then they could be covered (or partially offset) by utilizing a modest portion of the excess IMF income generated by the “Crockett” endowment. To encourage emerging market participation, a tiered interest cost offset structure could be agreed – whereby, a greater percentage or even all of their interest costs could be financed using gold income proceeds.
Selling or mobilizing pre-1978 gold as discussed previously would be politically difficult in the United States and elsewhere. However, the sale of post-1978 gold has already been approved by the IMF’s members (the Crockett endowment) to finance IMF surveillance and other non-income generating IMF activities as well as to help subsidize the IMF’s concessional lending to low-income countries. Since the approval of that sale, IMF income has risen due to recent global crisis lending responses, which has generated substantial income. Meanwhile the price of gold has remained higher than anticipated when the sale was approved. Both factors have combined to create a surplus of gold related revenues for the IMF. While the specific amount is difficult to project, it is likely that resources could be made available to cover such interest costs assuming the requisite political will.
In short, it would make sense to have the proposal include provisions to cover interest costs of all participating countries, perhaps in a form that permitted recourse to gold to finance the costs above some fixed interest rates, which could vary on a sliding scale as a function of participating members’ per capita income. At their core, pursuing some combination of SDRs and gold furthers the fundamental logic of utilizing existing global assets to finance programs that deal with global public goods and bads – of which, climate change is a particularly paradigmatic example.
In this paper, we have avoided discussion of specific institutional modalities. Agreement on any option for collective financing is unlikely, including in the case of the United States, without clarity on what new or existing institutions, under what governance and management arrangements, would deploy the resources. In the case of our proposed use of SDRs and gold, we should be clear that the IMF would not manage or control the third-party entity (nor has the IMF Managing Director or staff ever suggested they should), as that activity would take it far from its core mission. At the same time, we note that discussion of how to raise the revenue for global public goods, including climate mitigation (and adaptation) need not in principle await discussion of how to deploy those resources, and that the advantage of collective mechanisms of financing for global public goods is that it invites as well as being dependent on global cooperation on a global challenge.