1The welcoming words of Stanley Fischer towards Indonesia’s decision to float its currency (Fischer, 1997) sparked controversy over the International Monetary Fund’s (IMF’s) handling of the South East Asian crisis. Jeffrey Sachs, for instance, reignited the debate about the legitimacy of the IMF’s interventions, commenting that instead of “dousing the fire” of the financial crisis spreading through Asian countries, the IMF “screamed fire in the theatre” (Sachs, 1998). Concerns about legitimacy have been usually premised on the IMF’s unequal governance structure and its policy advice. Covered widely in several works, the IMF’s legitimacy issues are often explored through the prism of political influence over the use of its resources (Breen, 2013; Buira, 2005; Momani, 2004; Thacker, 1999).

4This paper examines how economic knowledge is embedded in policy templates and shows how expertise is shaped to resolve the IMF’s legitimation problems. Barnett and Finnemore (2004), for instance, draw on Porter (2001) to examine how expertise, by appearing objective, is a defence to the longstanding accusations towards the Fund’s politicised behaviour. Existing work on the politics of knowledge within the Fund, identifies the use of policy templates as a source of “cognitive authority” and “indirect political power” (Broome and Seabrooke, 2011, 9). Seabrooke assesses whether policy templates used to benchmark the financial sector are “tools of domination in the quest for transparency” or whether these templates allow for a “pragmatic use of numbers” to allow for “reasoning and sense-making” (2012, 490, 500, 493). This paper draws out the role of the public as an indispensable aspect of how IMF expertise is shaped to resolve legitimacy problems by taking as a case study the creation of the IMF’s Debt Sustainability Analysis (DSA) template.

5The IMF’s Debt Sustainability Analysis (DSA) is a template to facilitate the measurement of a country’s debt sustainability that was introduced in 2002 and 2003 (IMF, 2002a; 2003). The template provides a pre-set format for guiding the way debt sustainability will be assessed by staff economists. This is a crucial policy area for the Fund; current policy prescribes that the IMF can only lend to its members if their debt is deemed sustainable in the medium term, and if it is not, the IMF is precluded from lending until restructuring or further concessional financing is secured (IMF, 2019). This paper shows the history of this policy device as arising from an attempt to correct previous practice that was publicly derided for being unsound. The introduction of the template attempted to make practices comparable across countries and to introduce a rules-bound quantification of access to Fund resources. This paper shows that the actual theoretical underpinnings of policy however are less important than how the ‘public’ perceived them. The paper uses archival material from the Executive Board from 2002 and 2003 to show that a significant rationale for introducing the DSA was to guide and quantify its decision making over the loan-giving process in order to attempt to constrain large loans and enhance its legitimacy in the public domain. This would arise from the perception of soundness in policy design, in turn relying on a perception of underlying theoretical rigour. Collective authorship is significant in raising the status and credibility of economic reasoning within and outside the Board, as a variety of audiences, as imagined by the Board, are drawn into policy design. Staff documents are written and presented to enhance the credibility of IMF expertise to external publics, and at the same time, the role of the public acts as a discipling device on IMF Staff. The paper first provides a brief overview of the Fund and its legitimacy issues in Section 1. Section 2 introduces the new policy, lays down its theoretical rationale and illustrates the template’s use. Section 3 covers some of the Board’s difficulties in applying theory in a practical setting. Section 4 discusses the introduction of the template to aide its decision-making process, Section 5 discusses the template as an internal disciplining device for Fund staff, and Section 6 as a public communication tool.

1. Problematic Governance

7The IMF was initially set up in order to make its resources temporarily available to its members in need of balance of payments support (for a detailed account, see De Vries, 1987). Following the US suspension of gold convertibility in 1971 and the movement of countries off the adjustable peg system by 1973, the IMF’s role evolved into a more general financial crisis manager. The 1970s financial turmoil included the breakdown of Bretton Woods, but also the oil price and interest rate increases, which pushed much of the developing world into a debt crisis in the 1980s and a resulting ‘lost decade’ in development. The growing IMF’s involvement in crisis management was accompanied with far reaching conditionality programmes, which have since been the focus of most criticisms of the Fund’s influence on national policies. Such influence ranges from stipulating macro targets, for instance, on fiscal balances, to a perennially growing array of issues which include privatisation of public assets, liberalising trade and product markets, and a host of market reforms (for a review of Fund conditionality, see Babb and Buira, 2004; Kentikelenis, et al., 2016).

10Russia defaulted in 1998 amidst an IMF programme, and in 2001, the IMF decided to further increase Argentina’s already exceptionally large programme whilst the economy deteriorated amidst people’s seething fury. Economists in the Fund were divided given the whimsical credulity in the projected debt dynamics on which this decision was based (Blustein, 2006; IEO, 2004). These had assumed sustainability where none was forthcoming, as Argentina’s subsequent deposit freeze, default on its debt and abandonment of the fixed exchange indicated. The IMF’s reputational costs only multiplied.

2. The Template and its Theoretical Rationale

15The policy implementation of this theoretical rationale presented difficulties. The Fund’s measurement of debt sustainability was framed by the concerns of the Fund to provide its resources under adequate safeguards (i.e. conditionality) through its loan programmes and their repayment. The proposed policy template was to rest upon the debt dynamics equation to trace the future time path of the debt-to-GDP ratio over a specified time horizon. Initially, the suggested time horizon was for a minimum of five years, although this has since changed. The link to the theoretical rationale was that if the debt-to-GDP ratio was on a stable or declining path within the specific time horizon, the solvency criterion would be met.

18The first was to make explicit the underlying macroeconomic assumptions on which IMF policy were grounded, while not revealing the way they were made nor proposing to homogenise the way projections were made across country desks. The second was to evaluate those underlying assumptions, by testing the sensitivity of the projected path to them. Figure 2 illustrates sensitivity tests to the key parameters of the baseline. These tests are conducted by setting key parameters to historical averages, and by shocking the macro variables separately and combined, by sizes that have been differently calibrated over time. For instance, instead of using assumptions of the baseline, the assumption could be that policy remains unchanged, by setting values to historical averages (Figure 2, Line 1), or that the country experiences a one-time depreciation shock (Figure 2, Line 7). This was done with the purpose of identifying possible underlying optimism in the baseline projection.

3. The Academic-to-Policy Tension

22However, the application of a theoretical notion of solvency to a policy template raises more fundamental issues. Solvency is deemed a theoretically clear idea, but it is difficult to quantify and usefully measure. To operationalise the notion of debt sustainability and make it useful for the IMF’s loan programmes, practical assessments or “pragmatic” assessments, as the IMF calls them, are required. The reason why the IMF begrudges the abstract formal definition of debt sustainability, calling it “precise” but “unobservable” is that it is quite simply possible to postpone generating primary surpluses to cover initial levels of debt far into the future:

The theoretical concept of sustainability based on solvency is problematic because it does not impose specific constraints on debt and deficits at any point in time … any level of debt and deficits could be compatible with the present-value budget constraint (Akyuz, 2007, 3).

23The difficulty in translating the solvency criterion from academic to pragmatic terms is evident in IMF statements. For instance, in IMF (2002a, 3), judgments about debt sustainability are in fact about “whether a country’s debt can be serviced without an unrealistically large future correction in the balance of income and expenditure.” Further on, it reads:

It is difficult to assess the feasibility of the primary surplus consistent with debt sustainability without first specifying the tax and expenditure measures that would be needed to achieve it, and judging whether these measures are sustainable over time, both technically and politically (IMF, 2002a, 20).

24Far from theoretical clarity, the notion of debt sustainability is framed in terms of “social and political” feasibility of the stipulated adjustment plans (IMF, 2002a, 5). The evaluation of whether an adjustment path is considered to be too politically or socially unfeasible is wholly controversial, and the role of the Fund in determining it contested. Bar the litmus test of social implosion, the IMF does not come up with a means to judge whether a future correction will be unrealistic and hence to accurately assess debt sustainability. Debt sustainability in practice therefore is not determined by the economics of the present value budget constraint, but rather by the cautious balance of finding the adjustment path that lies just shy of preferring default and halting debt service:

Solvency needs to be viewed in relation to the adjustment path that is not only economically feasible, but also socially and politically acceptable such that default is not a preferred option (IMF, 2002a, 5).

25Being a creditor itself, often the only one willing to supply funds in a crisis granting it the status of preferred creditor over other creditors, undermines the IMF’s neutrality in measurement. It also reconfigures debt sustainability as a political notion grounded upon the conflict between creditors and debtors. If the debt-to-GDP path is on an upward or explosive trend, this would put pressure on the IMF to mediate with private creditors, who, if defaulted upon, would not receive full repayment. As there is no explicit rationale for the Fund’s view on whether an adjustment path is considered to be too politically or socially unfeasible, the other indication that it relies on is based on the comparison with the costs of a debt restructuring—choosing adjustment over collision with other creditors. In the words of the IMF:

• 20 In subsequent DSA templates, cross-country comparisons of size and rapidity of fiscal adjustment ar (...)

In principle, assessing whether bringing down debt ratios through a primary adjustment is too costly requires looking at the alternative by evaluating the costs of bringing down debt ratios through debt restructuring (IMF, 2011, 6).²⁰

4. Quantifying Exceptional Access to Fund Resources

In discussing the aforementioned criteria, Directors emphasized in particular the importance of rigorous debt sustainability analyses to support requests for exceptional access (Krueger, EBM/02/94).

29Board members strongly supported the introduction of a robust DSA as a qualification of the exceptional access criteria:

Debt sustainability assessments should play a crucial role in restricting exceptional access to truly exceptional cases, as well as in making it clear when required that drastic measures such as debt restructuring are necessary to help avoid prolonged Fund lending that is undesirable (Yagi and Miyoshi, EBM/02/94).

30The reasons for this were both to “strengthen accountability and ownership” but also to necessitate “a more systematic analysis of a country’s capacity to repay the Fund, which would be closely linked with debt sustainability analyses” (Lundsager and Baukol, EBM/02/94). If exceptional finance combined with domestic adjustment as stipulated in an IMF programme could not guarantee debt sustainability in the medium term with a high probability, then a debt restructuring would be needed. Bischofberger, the Executive Director for Germany, made it clear that if the sustainability condition is not met, then the onus ought to be on private creditors accepting a reorganisation of their claims. Hence, he joked that

while staff seem to assume that ‘Capital Account Crisis plus Assumed Debt Sustainability justifies Exceptional Access,’ we would argue that ‘Capital Account Crisis plus Uncertain Debt Sustainability requires Debt Restructuring’ (Bischofberger, GRAY/02/1222).

31Although the new methodology that was being developed at the Fund was deemed crucial in determining access levels and whether the repayment of the Fund was being jeopardised, there was no full faith in the newly developed methodology either. Some Directors noted that

we are afraid that we are still far away from having an adequate analysis of debt sustainability. The new methodology that was approved by the Board recently to assess sustainability (SM/02/166) is a step in the right direction, but it is still work in progress. (Oyarzábal and Beauregard, GRAY/02/1210).

32While the discussion about the introduction of the template emphasised the supremacy of judgement over strict quantification, the same could not be said for the rationale for introducing the template in the first place which was intricately tied into providing a means to quantify the exceptional access criteria requiring a rigorous analysis that showed that debt was sustainable with a high probability.

5. Disciplining the Staff Experts through Public Scrutiny

33The 2002 introduction of the template was meant to mitigate the adhocness and undue optimism with which some country desks pursued their assessments enabling programme approval without adequate economic rationale. Although the IMF tried to redress problems of credibility through greater transparency by making known on what basis decisions were made, the way macro variables were constructed remained concealed. This left ample discretion still available for each country desk to make projections and hence model the economy in whatever way it deemed appropriate, rather than imposing a common way to generate the underlying projections. Discussions at the Board swarm around the cumbersome fact that the Fund’s baseline is inaccurate, and Directors expressed concern that the IMF’s projections were not taken seriously in the eyes of the public:

Given that debt sustainability assessments are central to key Fund decisions in program contexts, … it is important that they be viewed as being based on solid judgments and are reasonably credible (Shaalan and Farid, GRAY/02/727; emphasis added).

34Board members appeared confused as to how those projections were crafted, and what their relationship is to WEO projections. One member explained:

the paper confirms that there is a clear bias towards over optimism in the projections, either because of a bias in the WEO projections, or because of a bias in the policy assumptions underlying the baseline scenario. If the WEO projections are biased towards over-optimistic debt projections, it is not only a problem for the debt sustainability assessments, but even more so for the credibility of the Fund’s projections in general (Egilsson, EBM/03/64).

35Faced with a general unwillingness to delve into the sources of possible systematic biases, the Board attempted to reduce them. One suggestion was “to include explanations in staff reports of why the most recent forecasts were faulty” (Jacklin and Dohlman, EBM/03/64). This was deemed “useful as a disciplining device, though…is not likely to eliminate systematic bias in projections” (Jacklin and Dohlman, EBM/03/64). Another suggestion was to include scenarios that would capture how staff projections differ from consensus forecasts (Martí, EBM 03/64). The problem “with over-optimism is not one of making mistakes, but rather that of not learning from them,” Directors noted (Padoan and Bossone, GRAY/02/730). As another Directors’ statement puts it: “The goal of the exercise is to [make] it more difficult to justify programs where there is insufficient adjustment or staff use unrealistic assumptions to make programs appear ‘sustainable’” (Lundsager and Ralyea, GRAY/02/734). The US Director, who represents the most powerful country within the Fund, explicitly argued that publication of the DSA outputs would enhance the “discipline imparted by the framework” (Lundsager and Ralyea, GRAY/02/734). Publication would mean that “there are improved incentives on the Executive Board not to support a program that appears to have a limited chance of success” (Callaghan GRAY/02/721). The publication of the template was therefore seen not only as a way to make the staff more accountable to the Board, but also, as the Members argued, to make the whole institution more accountable to whoever would be affected by its policy decisions: “sustainability assessment in the public domain is in everyone’s best interest, especially the country’s,” one Director claimed (Bennett GRAY/02/729).

36The role of a public was crucial—the Board seemed to believe that IMF expertise may be somehow disciplined by being made public. The IMF surreptitiously used the public to provide the appearance of rigour and to enhance its external credibility. “A program’s credibility may be enhanced if it is apparent that the Fund has considered a range of potential outcomes when making its decision” (Callaghan GRAY/02/721). Nevertheless, the IMF staff described the interpretation of the debt sustainability outputs as “still more of an art than a science, with a large element of judgment required” (IMF, 2003, 40). The art of the debt sustainability exercise far dominated any discussion of its scientific basis at the Board level.

6. Institutional Authorship Needs the Public

37The DSA was revised between 2002 and 2003 by introducing two new scenarios: a “no policy change” scenario and a historical scenario as well as an indication of the debt stabilising primary balance. The template was tailored so that the economic reasoning that arises can yield specific and desired results. For instance, the template output could yield a sustainable trajectory by being predicated on dramatic fiscal adjustments and optimistic assumptions about growth, implying less need for debt relief. As part of the dialogue with country authorities, the template outputs could be used to pressure a country into adopting policies to compress demand and reduce incomes. Were the DSA output to suggest unsustainability, it would mean that an IMF programme was ill-defined, as domestic adjustment and IMF financing could not guarantee repayment of IMF resources, and would require an alternative course of action. If debt cannot be made sustainable through adjustment, the pressure would be on creditors to accept a restructuring of their contracts.

38With this in mind, one of the Directors remarked: “Our goal should be to ensure that the authorities face the right incentives when confronted with very high debt burdens” (Bennett, EBM/03/64). When the Board discussed the addition of more scenario analysis, the rationale was that these “could provide more persuasive arguments for the need to adjust policies” (Portugal and Tombini, EBM/03/64). To achieve this, Directors remarked that the staff needed to explain the objective of the template, its mechanics, and proposed refinements more clearly. Communication is key for soliciting the desired changes, for making it more likely that the new template can leverage policy response:

Better technical understanding by the authorities should not only lead to improved scenario identification and analysis but also, and perhaps more importantly, enhanced ownership of the courses of action that such scenarios may prompt (Callaghan and Amador, EBM/03/64).

in a capital market crisis, what matters is not the Fund’s assessment, but what the markets think about sustainability. The best chance of the markets sharing the Fund’s view is if as much information as possible which underlines the decision is made available (Callaghan, GRAY/02/721).

41In particular, releasing the information could reduce market participants’ monitoring costs (Yagi and Toyama, GRAY/02/724). Publication was vociferously opposed by those Directors who believed that the operational value of the exercise was internal decision guidance about how to use the IMF’s resources and ascertain repayment prospects (Shaalan and Farid, EBM 02/7). Representing the views of several Directors, one mentioned: “The new framework does not in itself provide sufficient grounds to ‘elevate’ the status of our debt sustainability assessments by making them public” (Shaalan and Farid, GRAY/02/727). This was because of the confidential nature of the information that the IMF is privy to, itself a product of the “special” role of the IMF in the international financial architecture.

43Given the origins of the template, the discussion about publication was also related to the outcome of other Fund policy developments, specifically, the proposal by Krueger to establish a sovereign debt restructuring mechanism (SDRM). One director mentions that “the introduction of an SDRM would strengthen the case for publication” (Yagi and Toyama, GRAY/02/724). Had the discussion in the Fund about a SDRM been fruitful, the DSA may have been destined for a far more pivotal role; if a SDRM was in place, the analysis of the newly proposed DSA template “would indicate whether an application of a debtor country to activate the SDRM is justified” claimed Wijnholds (GRAY/02/726). In that case, DSA analysis could be used more directly to effect private sector involvement by indicating how much debt had to be reduced for it to be brought to levels deemed sustainable and what the expected contribution of private sector creditors to that reduction would be. The DSA was understood as important not only as a means to guide IMF policy, but to define the “need—and scope—of private sector involvement” (Von Kleist and Fabig, EBM 03/64). Though the idea for a SDRM was formally dropped in 2003 at the IMF Ministerial meeting, the life of the DSA continued with the focus on domestic contraction and surplus generation for debt repayment fortified.

7. Conclusion

44This paper showed the creation of the IMF’s Debt Sustainability Assessment template as the development of a socio-technical tool to guide decision-making of the IMF. The introduction of the template was part of a broad effort to redress policy blunders and improve appearance in the public sphere. The paper established that analytical issues are not free from political interpretation when put to operational use. The DSA embeds surplus generation instead of favourable debt dynamics for debt repayment, despite the ostensible academic clarity, the need for pragmatic definitions muddies the process of measuring debt sustainability in practice. The IMF relies on a pragmatic definition of debt sustainability that rests upon a notion of politically and socially feasible adjustment and a technical tool that provides little means to assess it. Instead, the key driver of establishing the DSA was the political decision to curtail access to extraordinary IMF loans through the introduction of a seemingly rigorous process to guide policy action. This reveals a tense amalgam of the technocratic and political dimensions of IMF work, overshadowed by an attempt to resolve credibility problems.

45The pre-formatted framework for analysis can be interpreted as a disciplinary activity to tame country desk teams to be more accountable for what they do. Introducing the template was part of an institutional curb to the tendency of producing biased analyses and a means to discipline IMF experts. In this respect, the theoretical underpinnings of the framework were less important than public perceptions in driving the analytical approach. The template attempted to resolve legitimacy problems by revealing the numbers on which decisions were made but provided little insight into how those numbers arose. This allowed continued room for reliance on overoptimism, and hence, less debt relief by creditors and placement of the burden of adjustment on the debtor. The use of collective authorship of staff documents enhanced the credibility of expertise and elicited desirable responses from the Board by framing requests for feedback in specific ways and developing scenarios to encourage policy response. There was not much on offer however as to how revealing numbers and introducing prefabricated routines for economists to fulfil could absolve the Fund from long standing legitimacy issues arising from its governance structure and nature of its policy advice.

This paper was prepared for the workshop on “Economics and Public Reason” hosted by the Centre Walras-Pareto for the History of Economic and Political Thought at the University of Lausanne in May 2018. I would like to thank the organisers and editors Harro Maas, Steven Medema, Marco Guidi, the participants of the workshop, and the members of the editorial team. The paper and earlier drafts benefited greatly from comments received by Cléo Chassonnery-Zaïgouche, James Forder, Martín Guzmán, Chris Marsh, Daniel Munevar, Mihalis Psalidopoulos, Francesco Sergi, and participants at the Center for the History of Political Economy at Duke University seminar. I warmly thank Beatrice Cherrier and Michele Alacevich. Financial support by the Swiss National Science Foundation is gratefully acknowledged.