Before standardization

Most forms of governance involve some conception of universality – of the things, values or principles that apply to everyone and not just to a few. Many IOs like the IMF or the World Bank view themselves as universal organizations, with virtually all states as members. They must therefore be careful to consider the universality of their principles and policies. Yet even those organizations that are less global in scope, like national aid agencies, would generally like their policies to be seen as a reflection of universal principles rather than particular national interests.⁷ Governance practices rely on universals in two primary ways. Organizations like the IFIs and donors seek to govern in the name of certain values that they deem to be universal – such as good governance, human rights, sound economics or accountability. At the same time, many of these organizations also govern through certain techniques or forms of expertise that they see as universally valid, such as particular economic theories or principles.

While IFIs during both the structural adjustment era and in more recent years have sought to govern in the name of and through certain universals, they have defined those universals very differently. As I will discuss in Chapter 6, the recent strategy of standardization, which underpins the good governance agenda and the standards and codes initiative, relies on a combination of moral and technical principles to justify its universality. This “moralization” of finance and development is reminiscent of a much earlier era: Robert McNamara’s war on poverty in the 1970s.⁸ For McNamara, the battle against poverty was a moral imperative, a set of “fundamental obligations accepted by civilized men” that defined the Bank under his leadership.⁹ Yet, not long after McNamara left the Bank in 1981, the organization underwent what Gerald Helleiner acerbically called “another change of religion,” rejecting its earlier emphasis on poverty and rediscovering the virtues of a trickle-down approach.¹⁰ Under the leadership of the former banker and new World Bank President A.W. Clausen, and more importantly under the intellectual direction of the economically conservative Chief Economist Anne Krueger, the Bank redefined its objectives in more narrowly economic terms, focusing on adjustment and efficiency.¹¹ In contrast to the 1970s, the 1980s/early 1990s was an era in which global economic leaders went out of their way to deny moral universals. Some Bank staff even went as far as attacking the moralizing tone of the McNamara years as “imposing foreign concepts of morality” on developing economies.¹²

Although the 1980s was therefore not a decade characterized by much explicit moralizing rhetoric, it was nonetheless underpinned by a set of universalist economic assumptions; these universals simply took technical rather than moral form. In fact, as poverty and even growth dropped down the list of priorities and economic adjustment came to the forefront, Bank staff’s approach to development arguably became even more universalist, as their toleration for policy diversity declined.¹³ The 1981 Berg Report, an influential Bank-funded analysis of development policy in sub-Saharan Africa, argued for the universal applicability of certain policies, including more export-orientation, a smaller public sector and more agriculture-friendly policies.¹⁴ Where McNamara had argued that the mix between public and private ownership of key industries was a matter of domestic choice, Berg and Krueger insisted that public ownership was inherently inefficient, and argued for the superiority of private-based alternatives.¹⁵ The Bank’s doctrine became increasingly rigid, internal debate was stifled and the message carefully controlled.¹⁶

Although the IMF underwent its share of organizational convolutions in the 1970s as the fixed exchange rate regime that it had overseen fell apart, it did not experience a doctrinal volte-face in its relationship with poorer countries like the World Bank. Instead, its policies underwent a more gradual series of changes from the early 1980s onwards. In some ways, the World Bank’s rediscovery of trickle-down economics and its embrace of neoclassical principles brought it closer to the path that the IMF had been on for quite some time. The IMF’s approach to adjustment had always relied on narrowly economic tools – chief among them the Polak model of monetary adjustment.¹⁷ The rigidity of this economic approach to policy had been attenuated by two crucial factors. The fact that IMF programs were generally of short duration, primarily aimed at balance of payments adjustment and designed primarily for middle-income and industrialized economies, meant that a simple set of principles could be reasonably effective.¹⁸,¹⁹ At the same time, the universality of the rules was always complicated by a pragmatic approach to their application, which allowed for more variation in practice.²⁰

During the 1980s, both of these compensating factors were undermined, creating a more universalist approach to economic policy. The IMF scaled up its lending to low-income countries (LICs) that often had more complex economic situations. To address these challenges, both the IMF’s informal conditions (letters of intent, ex-ante conditions and the newly created structural benchmarks) and its formal performance criteria grew in scope and number.²¹ As supply-side economics became increasingly influential at the IMF, what had been a relatively narrow set of policy tools began to expand to cover other economic issues.²² Both Bank and IMF staff began to target a much wider array of domestic policies including trade liberalization, tax reform and eventually privatization, in the hopes that these more “structural” forms of adjustment would yield longer-term economic stability.²³ And although Fund staff and the Executive Board retained their pragmatic approach to interpreting the conditionality guidelines, they did so increasingly to enable the expansion of conditionality into these new areas.²⁴,²⁵

This gradual but relentless expansion in conditions at both the IMF and the World Bank was underpinned by economic assumptions that were believed to be universally applicable. John Williamson famously labelled them the “Washington Consensus” – a set of policy prescriptions based on a combination of neoclassical and supply-side economic assumptions that was believed to provide a universal recipe for economic reform.²⁶ It was in the name of these economic universals that the institutions sought to justify their policies in the 1980s. Unlike McNamara’s attack on poverty a decade earlier, and efforts to “civilize globalization” several decades later, these universals were not articulated in moral terms (although it was of course implicitly normative in its distinction between good and bad economic policies). Instead, trumpeting the wonders of “efficiency,” the IMF and World Bank sought to justify their increasingly interventionist and controversial policies through a language of technical universality. At the same time, the universal principles through which IFI actors sought to do the work of governing were highly rigid: they were exclusively economic and took the form of absolute rules, rather than the broader and more flexible standards that have become influential in recent years.

Before ownership

This rigid technical approach to finance and development also had significant implications for the ways in which the two organizations dealt with the problems of politics and particularity – the chief concerns that the later strategy of country ownership has sought to address. IOs, bilateral donors and NGOs have always confronted the challenge of balancing their claims to universality with a need to respond to different countries and contexts. Over the past decade and a half, this attention to particularity and politics has become a very visible part of the IFIs’ rhetoric and policies, as they have focused on ownership as the key determinant of policy success. Yet the concept of country ownership has only become influential since the mid-1990s. In earlier times, the Bank and Fund relied on rather different approaches to address the challenges of domestic politics.

In the 1970s, both organizations tended to rely on a strategy of separating politics from economics, treating “political” issues as the domestic issues to be decided by borrowing governments and “economic” issues as more universal in scope and therefore fair game for IFI action.²⁷ Thus McNamara made a clear distinction between what he called economic and political human rights, arguing that the Bank could be active in promoting the first of these, but that it could not get involved in the second.²⁸ At the IMF, there was a similar effort to separate politics from economics. In their first formal debate on conditionality guidelines in 1968, for example, staff and many Board members argued that one of the ways of ensuring that the IMF respected member states’ political priorities and values was by avoiding imposing conditions on a borrowing country’s fiscal policy (e.g. taxation policies and the budget balance), and focusing instead exclusively on monetary policy (e.g. targeting interest and exchange rates).²⁹ A Fund staff report to the Board noted that:

The staff thus sought to develop programs that were seen to be as apolitical as possible, and saw their limited focus on monetary policy as a way of ensuring this.³¹

In the 1980s, as both institutions expanded into ever-greater areas of their member countries’ policies, it became difficult to sustain such claims about the clear lines separating politics from economics. While the organizations continued to maintain that their policies remained apolitical, they did so increasingly by redrawing the boundaries between politics and economics. Gradually, more and more aspects of a state’s activities came to be viewed as economic problems. IMF staff, for example, quickly overcame their hesitation about the political overtones of fiscal conditionality, first allowing conditions on fiscal policy, and then moving onto more structural reforms. The Bank also saw a gradual expansion in both the number and scope of its conditions. While the first SAL, to Senegal in 1980, included thirty-two conditions, by 1990 the average number had risen to fifty-six.³² Structural adjustment conditions initially focused on balance-of-payments deficits, seeking to reduce them through export promotion or budget deficit reduction.³³ By the mid-1980s, however, the programs began to focus on supply-side and microeconomic issues, including prices, taxes, financial regulations, privatization and labour market policies: all of these issues, which not long ago would have been seen as matters of domestic political choice, were now seen as primarily economic.³⁴

This shift was widespread. As Gerald Helleiner noted, even Elliot Berg himself, the author of the influential 1981 World Bank report on sub-Saharan Africa, went from suggesting in 1963 that it was “not the business of outsiders . . . to quarrel about the suitability of the goals set out by socialists in Africa,” to arguing that an export-oriented economy and a smaller public sector would benefit just about any African economy.³⁵ While the Berg Report did note that some of the proposed reforms were likely to be “politically thorny,” and recognized the fragile political context in many African states, one of the central messages of the report was the importance of African countries paying less attention to “political consolidation” and more to “the efficiency of resource use.”³⁶ This conclusion, in stark contrast to the more recent consensus that political stability and institutional capacity are vital to the success of development programs, points to the tendency of IFI staff to deny the political character of their new interventions even as they moved into new, more fraught domains.

The growing faith in the universality of certain neoclassical economic principles allowed both the IMF and the World Bank to pay less attention to the particular situations faced by individual countries. Although directors from developing countries did raise the issues of domestic political constraints and urged the IFIs to respond, the fact that politics were treated as a separate domain meant that they were rarely taken seriously or integrated directly into policy. The only real place for addressing particular political contexts was in the form of exceptions or exercises of judgment.

One of the dominant strategies of both the Bank and the Fund during the structural adjustment years was thus to colonize new terrain as economic and therefore subject to universal economic principles rather than particular political values. Although IFI actors did recognize that domestic political constraints could be a source of policy failure, it was not a problem that they could address directly, since the political was still viewed as beyond the pale of IFI expert practice.

Before risk and vulnerability

This tendency to bracket politics or to treat it as an economic issue was also connected to and sustained by a conception of time that focused on the short-term, assumed considerable continuity between present and future, and treated shocks as exceptional events. This conception of the temporality of policies was quite different from the organizations’ more recent preoccupation with risk, which involves more attention to the ways in which policies evolve over time, and to the unpredictability of the future. As I have discussed elsewhere, all organizations must find some way of coming to terms with various unknowns.³⁷ International organizations operate in an uncertain environment, in which the long-term success of their policies depends on factors that are beyond their control. This uncertainty has become a more serious preoccupation of IFIs and donors in recent years than it was during the structural adjustment era. The 1980s’ more confident conception of the future was underpinned in part by the narrower focus – and tidier ontology – of IMF and World Bank policies: by denying the messy complexities of the political and social character of economic adjustment and development, focusing on a shorter time horizon and treating shocks as isolated events, staff could ignore many of the complicating factors that might upset their programs’ evolution in the future.

From its inception, the IMF was created specifically to address short-term balance-of-payments problems: in other words, to provide temporary financing to enable member countries to adjust their economy enough to bring their exports back up and into balance with their imports.³⁸ In this respect at least, the organization did respect the goals of one of its architects, John Maynard Keynes, who suggested that the IMF should be a kind of “clearing union” (rather like a modern-day credit union) in which countries could obtain temporary overdrafts (or credit lines) when in need.³⁹ As the convention of “stand-by arrangements” evolved from the 1950s onwards, countries were able to negotiate access to financing for between one and three years. Following the Polak model, these programs sought to act quickly on borrowing economies by limiting budget deficits and credit creation, rather than tackling longer-term challenges.⁴⁰ Thus, as James Boughton points out, until the creation of the extended financing facilities in the mid-1970s, IMF programs were of a short enough duration that its staff and directors did not need to think through possible tensions between adjustment and growth or consider the long-term sustainability of their prescribed reforms.⁴¹

The World Bank has always had a longer time horizon than the Fund, given that its mandate was initially to help reconstruct Europe after the Second World War, and thus focused on development rather than on short-term adjustment. Yet in the 1970s and 1980s that time horizon was still quite limited in comparison with present practice. The organization’s early emphasis on projects was consciously “bounded” in both time and scope, requiring little thought about the longer-term effects of development efforts.⁴² Even when the Bank first developed SALs in the early 1980s, it was assumed that this kind of economic reform would be a short-term “big bang” rather than a more gradual, long-term process. Bank staff did recognize that adjustment was painful, often requiring the elimination of certain subsidies, dramatic changes in interest and exchange rates, or the liberalization of trade policy – all of which could be politically as well as economically disruptive. Yet it was hoped that this disruption would be brief, with the economy moving back into a growth-oriented phase shortly after adjustment. Underlying this short-term approach were some basic neoclassical economic assumptions: although it was generally accepted that the best economy is one without distortions (understood as government constraints on a free market), it was believed that if you removed just some of the distortions, the remaining ones could have unintended consequences, leading to further deterioration.⁴³ This so-called problem of the second-best solution led economists and policymakers to opt instead for an “all at once” approach to economic liberalization.

Over time, some serious strains did begin to appear in the IMF and World Bank’s optimistic time horizons: as their policies became increasingly ambitious and wide-ranging, and it became clear that adjustment was not working as quickly as had been hoped, both organizations were forced to conceptualize their policies over an increasingly extended time horizon. The global economic system was also becoming more uncertain in the 1970s and 1980s, once the fixed exchange rate system collapsed and the Organization of Petroleum Exporting Countries (OPEC) crises took their toll, forcing both IFIs to address the effects of shocks. The IFIs responded by stretching out the existing set of temporal assumptions rather than reconceptualizing them entirely; staff continued to assume that policies would work as expected, but just take longer than originally anticipated, and with a few extra bumps along the way.

By the mid-1970s, it became clear to IMF staff and directors that short-term loans did not always provide enough time for borrowing members – particularly poorer countries – to make the changes necessary to turn things around. Developing members were particularly supportive of the idea of a longer-term facility; after some deliberation, the Extended Fund Facility (EFF) was created in 1974.⁴⁴ This was followed in the 1980s by the creation of the Structural Adjustment Fund (SAF) and the Enhance Structural Adjustment Fund (ESAF), both of which sought to provide financing over a longer time in order to enable more profound structural economic reforms. Not long after the Fund’s staff created such extended facilities, they became concerned about what they termed “prolonged use” of the IMF’s resources. By 1987, over thirty developing members were considered prolonged users, given the number of times that they had borrowed from the Fund and the outstanding credit that they owed.⁴⁵ As the organization lent more to developing countries and placed more emphasis on structural reforms, it therefore also began to conceptualize the duration of policies over a longer time frame.

IMF Directors also debated and ultimately created a number of facilities designed to respond to potential shocks. The EFF was designed in part to respond to the disruptions caused by the first OPEC crisis.⁴⁶ The Fund also set up two different oil facilities designed specifically to help countries whose balance of payments difficulties were caused by the increase in oil prices and, starting in the 1960s, the IMF also provided emergency disaster relief in certain cases.⁴⁷,⁴⁸ These facilities recognized the potential for unforeseen shocks to upset economic development, but they did so by treating the problems as very specific and generally isolated events.

At the World Bank, staff found themselves facing some similar dilemmas. Despite their optimistic assumptions about the speed with which structural adjustment policies would take hold, it was becoming increasingly evident that the complex changes they were trying to achieve were taking longer than anticipated.⁴⁹ In response, Bank staff began to extend the time horizon of their policies from the mid-1980s onwards, sometimes distinguishing between quicker “first” and slower “second-generation” reforms.⁵⁰ Towards the end of the 1980s, the Bank staff began to restructure their evaluation processes to try to take account of the longevity of their programs’ effects. In 1985, the World Bank’s Operations Evaluation Department (OED) introduced the category of “sustainability” in its project evaluations, which it defines as the likelihood of a project sustaining its benefits after completion.⁵¹ The Bank also began expanding its provision of emergency loans for countries whose balance-of-payments difficulties stemmed from various natural shocks, such as the drought that hit Ethiopia and Sudan in the mid-1980s.⁵² In 1984, the Board adopted guidelines on reconstruction, later formalized into an operational directive on “Emergency Recovery Assistance” in 1989.⁵³ Yet, as Kapur et al. note, both McNamara and Clausen were resistant to this kind of lending, seeing it as a form of “relief” rather than economic “reconstruction.”⁵⁴

Although the Fund and the World Bank thus conceptualized and operationalized the role of time differently, they both found themselves having to integrate a longer time horizon into their policies throughout the 1980s, as they deepened and expanded the structural aspects of their programs. They also began to find ways of addressing the unknown through peripheral policies designed to address certain kinds of shocks. Yet they both continued to see the future as relatively linear and to treat contingency as a matter of isolated shocks rather than as a more profound challenge to policies’ success.

Before results-based measurement

If there is one defining feature of modern bureaucratic organizations, it is probably the drive to count, measure and evaluate – in short, to translate the world into numbers. As I suggested in the previous chapter, the expert authority that is so crucial to institutional survival relies heavily on such forms of measurement and calculation. Yet, while the urge to quantify has remained constant among IFIs and donors, the form that these measurement practices have taken has changed over time. In fact, processes of evaluation and measurement have become an increasingly important and visible part of the governance of development finance over the past decade. Not only are key IOs and donors paying more attention to their own measurement practices, but they are also pressuring developing countries to adopt new evaluation strategies. The central concept around which many of these new governance practices hinge is that of results.

Although there is a long history of program evaluation at these organizations, particularly at the World Bank, these earlier practices of measurement were quite different from the current emphasis on results. As I discussed above, the policies of the 1970s and 1980s were based on universalist economic principles, with relatively short time horizons. Although, of course, the effects of these policies were far from straightforward, they were nonetheless conceptualized and operationalized in relatively simplistic terms. The metrics used for evaluating the policies, and the definitions of “successful” and “unsuccessful” programs, were also relatively straightforward and short term, focusing, for example, on the economic return to projects. At the same time, however, even in the earliest days of evaluation, IFI staff were aware of some of the dilemmas of expertise, recognizing the difficulties of adequately measuring the more complex dimensions of development financing.

The World Bank’s institutional fascination with measurement and evaluation can be traced to the influence of Robert McNamara, who was known for this love of numbers. He put it particularly clearly when he once remarked: “I see quantification as a language to add precision to reasoning about the world.”⁵⁵ In his efforts to expand Bank lending and reduce poverty in the 1970s, McNamara introduced a host of measurements and targets for tracking loan volumes and poverty statistics. He also created the OED in 1973, which published annual reports evaluating Bank projects, a practice that continues today.⁵⁶

These early reports assessed Bank projects as “successful” or “unsuccessful.” They did so by applying a straightforward metric based on the volume of lending – a measure of policy inputs – and on the rate of economic return – a relatively narrow measure of output.⁵⁷ Fund staff focused almost exclusively on ex-post evaluation, “assessing the results of work already done, rather than trying to second-guess the decisions of those responsible for current work,” in the words of Christopher Willoughby, Director of the OED in the early 1970s.⁵⁸ Yet Bank staff and leaders had already begun to recognize the limits of their measurement efforts. McNamara himself made it very clear that he saw discrete projects as means to the end of policy influence: for him, the ultimate goal of the Bank’s efforts was to effect the kinds of changes that would ultimately achieve broad-based development and reduce poverty.⁵⁹ Yet he and the OED staff were unable to find a way of measuring that influence.

By the 1980s, the Bank staff had developed techniques to enable them to formalize this effort to influence borrowing countries’ policies through structural adjustment loans, but they were still struggling with the problem of measuring their effects. Most of the OED’s annual reports continued to focus on projects (rather than policy loans) and to define success primarily in terms of project effectiveness (ability to attain initial objectives) and process efficiency (which takes into account some factors that may have complicated the original goals).⁶⁰ By these relatively narrow measures, Bank programs were quite successful, with most years finding success rates of between 80 per cent and 85 per cent – with the notable exception of sub-Saharan Africa, which continued to suffer from much lower rates of success.⁶¹

At the same time as they focused on these narrower determinants of policy success, it is clear from these documents that Bank staff were struggling with the question of how to measure the less tangible but important aspects of their loan programs. As I noted above, in an effort to capture the longer time horizon of structural adjustment initiatives, OED staff introduced the idea of sustainability in 1985 in an attempt to measure the likelihood that projects would sustain their benefits after Bank funds had been disbursed.⁶² The evaluations of sustainability consistently found a lower rate of success than the overall assessment, which tended to focus on the project up to the moment of completion.⁶³ As they began to try to assess such complex factors, OED staff expressed some frustration that the measures remained subjective and difficult to quantify.⁶⁴ Moreover, as staff noted in their 1990 review, although most projects could be evaluated through a cost-benefit assessment of their economic rate of return, they had to rely on their subjective judgment when evaluating the success of structural adjustment programs.⁶⁵ Bank staff had not yet invented a way of adequately translating these more complex problems in quantitative terms – a challenge that would lie at the heart of the later turn to results measurement.

The IMF’s history of evaluation has been more sporadic than the World Bank’s.⁶⁶ The Fund did not create an independent evaluation office until 2001, and generally relied on occasional internal staff evaluations (and the odd external review) until that time. The IMF Executive Board did, however, mandate periodic evaluations of the conditionality policy beginning in 1979.⁶⁷ The Fund’s initial reviews focused on formal performance criteria, measuring the degree of compliance with these conditions and assessing the attainment of key objectives, including inflation rates, balance-of-payments changes and growth rates.⁶⁸ Staff also began to use a variety of methodologies in determining program success.⁶⁹

What is striking about these reviews is the fact that they consistently found very low levels of “success” in conditional Fund programs. The 1979 and 1981 reviews found that the achievement of most of the performance criteria objectives was “mixed” at best, while the 1982 report found that performance “fell short of expectations in many cases.”⁷⁰ In fact, as IMF historian James Boughton summed up the cumulative results of the evaluations in the 1980s, the success rate was somewhere between one quarter and one half of the programs initiated.⁷¹ Yet, while these relatively low levels of success were a matter of concern, they did not cause the organization to significantly redesign its approach. Instead, both the staff and most of the Board concluded that the problem was not the model of adjustment itself but a wide range of exogenous problems,⁷² including the fact that certain objectives, like growth, took longer to achieve than was allowed for in the evaluation time frame,⁷³ the role of domestic factors including a “lack of political commitment,”⁷⁴ and the effects of external shocks.⁷⁵

Like their peers at the World Bank, IMF staff were also aware of the limitations of their efforts to measure the success of adjustment programs. In their 1979 report they noted the “significant element of judgment” involved in any evaluation and raised some important questions about the difficulty of establishing a clear causal connection between IMF policy instruments and final outcomes in inflation or growth rates.⁷⁶ In their 1979 and 1981 reports, they also noted that the time frame of their evaluations was too short to assess the implications of policies that take longer to take effect.⁷⁷

The cultures of evaluation within the Fund and the World Bank in the 1970s and 1980s were clearly very different from one another. The Bank had adopted a systematic approach to evaluation relatively early on, one focused primarily on simpler metrics such as inputs and easily quantifiable outputs, but which was gradually beginning to tackle more complex metrics like sustainability. The IMF, on the other hand, adopted a more ad hoc approach to evaluation and continued to be remarkably sanguine about what appeared to be much lower success rates. Both organizations contented themselves for the most part with relatively straightforward metrics for evaluating policy success and failure. At the same time, both struggled with the limits of their measurement techniques and began to experiment with different ways of measuring less easily quantifiable aspects of their programs.

The fragility of expert authority

Why was the IFIs’ expert authority attacked and, more importantly, undermined? After all, we have generally come to think of expertise as not only the most pervasive but also the most secure basis for authority. We owe this perspective on technocratic authority to Max Weber above all others, for his powerful depiction of modernity as subject to the progressive colonization by the technical-rational authority beloved of bureaucracies. Yet, it is actually through a particular reading of Weber, by the political theorist Sheldon Wolin, that we can also begin to grasp the fragility of expertise. Wolin suggests that if one reads Weber’s political theory and methodological work together, it becomes clear that

So far so good for our IFI actors, who rely so heavily on the methodological certainties of expert authority. Yet Wolin suggests this scientific solution to the problem of authority is only ever temporary: even the fact–value distinction that was at the heart of Weber’s methodology was an article of faith. It had to exist in order to ensure that values remained within the realm of choice.¹⁵ In such a world, the methodologist, like the Calvinist in The Protestant Ethic, or the charismatic leader in Economy and Society, is a heroically moral figure, who must not only have faith but actively foster it at times when belief flags.¹⁶,¹⁷,¹⁸ Weber himself lived through such a moment, during the German methodological debates, “when the nature of the social sciences qua science was being contested.”¹⁹ For Weber, Wolin suggests,

Methodological crises are thus generated by ontological contingency. At moments of crisis, when foundational values no longer seem to fit the changing world, methodologists must act: not only finding new methods, but also grounding them in new epistemological claims. They must challenge out-of-date modes of analysis and establish new ones, thus rebuilding the foundation – and restoring the faith – that makes social scientific analysis possible.

There is much in Wolin’s description of potential crises of expert authority that resonates with the experience of development finance actors over the past two decades. We have witnessed the rise of new methodological debates that have sought to challenge and replace the old foundations of development knowledge. Moreover, as I will discuss below, some of the most vigorous and significant debates have been focused on questions about how to measure success and failure, while new policies that have emerged have sought to measure new things in new ways: measuring risk, results, ownership and best practices, rather than compliance with conditions. This is not simply an example of ideational change. It is not just how people frame the world that has changed, but also how they count and calculate and seek to engage with things; a methodological, and ultimately an epistemological, transformation has been underway.

We face methodological and epistemological limits because the finitude of our frameworks and metrics must come face to face with the open-ended character of the world. As Weber points out, the world is “perpetually in flux,” posing a constant challenge to our efforts to understand it. Failure, in this sense, is built into all of our efforts to understand and transform a world that resists us. It is a central feature of modern theory and practice. We do not fail simply because we have not recognized the changes that have occurred, as in the classic dilemma of always fighting the last war. We also fail because the contingencies of the world force us to change our metrics and redefine what counts as success and failure. Failures are always contested, as various actors define or deny them in their own ways. Yet some failures are so contested that they raise these more fundamental methodological and epistemological questions about what counts as failure.

In such moments of more profound problematization, we need to rethink basic categories and re-engineer our practices. In the process, we make the tacit background of our everyday lives the subject of reflexive thought and debate, at least for a time, until we re-establish our methodological foundations and forget their fragility. While such gaps between the world and our efforts to make sense of it will always appear eventually, they can also be intensified or accelerated under certain circumstances. In the case examined here, both the IMF and the World Bank began in the 1980s to delve into increasingly complex arenas, as they moved into structural adjustment lending. Yet the methodological categories through which they sought to make sense of these more complex objects remained narrow and simplistic, increasing the tensions in their claims to expert authority.

What happens when such gaps grow wider and the fragility of expert claims is exposed? Two other theorists also provide us with some additional clues to such moments. Andrew Barry suggests that agreed practices of measurement and calculation can act to reduce political dispute, by fixing certain decisions and excluding them from the debate; so, for example, if we agree that a development program’s success can be measured using a certain set of metrics, then the assumptions underlying those choices are not subject to dispute.²¹ Yet Barry also points out that such depoliticizing effects are not guaranteed: systems of calculation and measurement rely on processes of standardization which are necessarily imperfect when faced with the complexity of the world.²² They are therefore inherently fragile, and can themselves be subject to political debate – producing something rather like the kind of methodological crisis that Weber was concerned about – and the kind of problematization that I have pointed towards.

Michel Callon uses the term “cold negotiations” to describe those debates in which the basic parameters of measurement are agreed, and “hot negotiations” to describe those in which the basis of calculations may themselves be subject to debate.²³ Many of the debates and transformations that I will survey in this chapter were effectively either hot or warm negotiations, in which the basis of calculation was itself up for grabs.²⁴ The debates surrounding discussions of the success and failure of development and adjustment clearly constitute this kind of hot negotiation. These debates about what counts as “aid effectiveness” have in turn informed other discussions about the importance of ownership and good governance, the need to manage risks more effectively, and the kinds of measurement techniques required. As I will discuss in later chapters, while many of these discussions were first either relatively cold or warm, over time much of the debate shifted precisely to the question of what counted as ownership, good governance, risk and meaningful measures of success and failure.

The politics of failure

What provokes these more profound debates? Although there are many potential causes, such hot negotiations often emerge in the context of highly contested failures. These are publicly visible failures that are seen as particularly serious and important, but over which there is significant disagreement about their causes and implications. Such contested failures often pose serious challenges for the actors or organizations that are seen as responsible, and can lead to new problematizations and new strategies. While such contested failures do not cause institutional change outright, they can unsettle taken-for-granted assumptions and practices, come to symbolize existing tensions, and help accelerate the processes of erosion and problematization. In the case of the IMF and World Bank, three contested failures in particular had formative effects on IFIs’ thinking and practices: the Asian financial crisis, the crisis of development in sub-Saharan Africa in the 1990s, and the more recent 2008 global financial crisis. These failures played important roles in the transformation of IMF and World Bank governance strategies – not so much because of what they objectively revealed, but rather because of how they were taken up and represented within and outside the organizations.²⁵

The 1997–8 Asian financial crisis was a serious and publicly visible failure of the international financial system. Yet it was also a very contested failure. On the one hand, critics, including the World Bank Chief Economist, Joseph Stiglitz, argued that the crisis and the IMF’s response to it were both failures of IMF orthodoxy: it was because Asian countries went too far in adopting the IFIs’ prescriptions of financial liberalization that they were left without the tools necessary to respond effectively.²⁶ Yet IMF staff and management saw the crisis as a different kind of failure – one with domestic political and institutional causes. Together with the US Secretary of the Treasury and mainstream economists, IMF staff argued that the domestic economies of the Asian countries were structurally unsound and distorted by “crony capitalism.”²⁷ Drawing on the increasingly influential small “i” ideas of institutionalist economics, they argued for the importance of reforming not just economic policies but also economic, legal and political institutions. This reading of the crisis meant that the Fund was not only justified but also required to expand its mandate, and encourage borrowing countries to undertake more profound kinds of institutional reform.²⁸

Another important failure that became a focal point for debate around the same time was the recognition of the “decades of despair” (the 1980s and 1990s) in sub-Saharan Africa.²⁹ In the 1980s, investment declined in the region, exports fell and real per-capita income and food production both dropped, while African governments took on ever-greater volumes of debt.³⁰ For the Bank in particular, the persistence of poverty in the region and failing to achieve sustainable development was a source of shame. No matter how brightly they might paint their reports on global and regional development outcomes, the fate of sub-Saharan Africa remained a dark stain. Yet, as with the Asian crisis, the question of what kind of failure this represented was itself the subject of contestation. For critics, it was a clear indictment of the World Bank’s heavy-handed structural adjustment policies. For many Bank staff, it was seen as a different kind of failure: above all, the lesson drawn from this experience was that domestic factors, particularly political capacity and institutional development, played a crucial role in determining the success or failure of development programs.³¹

This preoccupation with the failure of development efforts in sub-Saharan Africa was not new. OED evaluation reports throughout the 1980s noted that the projects in sub-Saharan Africa had consistently high rates of failure in comparison with other regions – without precipitating the kind of radical rethinking that began in the 1990s.³² As my brief overview of the World Bank’s history in the previous chapter reveals, the 1981 Berg Report had also focused on the region’s difficulties and emphasized the importance of domestic factors – yet it drew rather different conclusions: the report downplayed the importance of achieving political consolidation and focused instead on structural economic issues, justifying the structural adjustment approach to economic governance.³³

More recently, another major crisis has had a destabilizing effect on the IMF, World Bank and donors: the global financial crisis that began in 2007. Whereas the Asian crisis was largely blamed on Asian domestic governments, it was simply not possible to blame this more recent crisis on other countries or on governments alone. Mainstream economists and IFI leaders finally began to see the markets themselves as a source of considerable instability – a sign of the failure of the West to adequately regulate financial practices and to anticipate the potential for devastating shocks. The recent financial crisis was also a contested failure, with critics arguing that it pointed to a profound failure in the global financial system, and IFIs, most Western leaders and many economists suggesting that the failures were more modest, requiring less radical changes to the system. Yet IFIs and donors did conclude that they had to pay more attention to the fundamentally volatile and contingent character of the global economic system. This more recent contested failure has therefore played an important role in precipitating a shift in how mainstream economists and IFI and donor staff conceptualize the world around them, leading them to place greater emphasis on risk, vulnerability and the ever-present possibility of shocks.

These failures precipitated debates not only within the organizations themselves but also among state leaders, non-governmental organizations (NGOs) and academics. British and Nordic country leaders seized the opportunity presented to pressure the IFIs to adopt the “aid effectiveness” agenda. In the US, during the final years of the Clinton Administration, Congress was extremely critical of the IMF and World Bank, with the Meltzer Commission proposing a reduction in the role of both organizations. What is most interesting for our analysis is not necessarily the IMF’s blunders in Asia, the continued poverty in Africa, or the growing global financial instability, but how and why these failures became important when they did, sparked particular debates, and helped foster new governance practices.

The problems (and possibilities) of politics

Politics always poses a challenge to bureaucratic institutions’ expert authority, given their claim of neutrality and objectivity.³⁴ Of course, this claim is always something of a lie, since even the most technical of operations has political implications. Organizations must therefore carefully navigate these tensions. During the structural adjustment era, the IFIs’ claims to expert authority depended in part on their ability to redraw the boundary between the political and the economic, redefining issues that had been deemed political as purely technical and economic. Where they did explicitly recognize the role of politics – usually as a problem – they rarely sought to tackle it directly, seeing it as beyond their mandate.

Yet the IMF and World Bank could not ignore politics forever. The deeper both institutions moved into the minutiae of domestic policies and practices throughout the 1980s and 1990s – imposing conditions on public pensions, price controls and privatization – the more vulnerable they became to charges of political interference. Their actions thus ultimately helped to fuel the problematization of the political dimensions of development finance. NGOs became increasingly vocal critics of the World Bank and IMF’s heavy-handedness, charging them with political interference.³⁵ The World Bank was the first to respond, and tried to defuse NGO criticisms through various outreach programs. For most of the 1990s, the IMF largely ignored its critics, as staff and management believed that part of their role was to have the “broad shoulders” needed to take the criticisms of domestic forces when a government instituted painful adjustment policies. The fallout from the Asian crisis changed all that, however, as criticism became damaging enough that the organization began to take it seriously.³⁶

Interestingly, although the two organizations’ increasing movement into domestic politics was the source of much tension, one of the ways that the staff in both organizations ultimately resolved it was by admitting, and justifying their attention to, domestic issues, rather than by continuing to deny that they were political. Although IMF staff remained coyer than those at the Bank about admitting the political dimensions of their policies, both institutions gradually found ways of tackling more political questions, as did donors such as the United Kingdom’s Department for International Development (DFID), the United States’ Agency for International Development (USAID) and the Millennium Challenge Corporation (MCC).³⁷ They did so in part by drawing on public choice theory and new institutionalist economics, both of which recognize the role of political pressure and institutional dynamics in economic adjustment, making them amenable to economic analysis. As long as it could be shown that a political issue had significant economic consequences, then it was fair game. “Political economy” (defined in public choice terms) became the preferred lens and euphemism for the previously forbidden subject of politics: one Vice President of the World Bank’s poverty reduction and economic management (PREM) network, for example, refused to let the staff hold a seminar on politics, but would let them hold one on political economy.³⁸

IMF and World Bank staff did not just start focusing more on overtly political problems, such as institutional reform; they also became increasingly interested in integrating political techniques into their governance strategies. Although the idea of participatory development in particular had been quite influential among NGOs and certain World Bank units (particularly the Social Development Group), it was only in the mid-1990s that participation was seen as a technique that could be integrated into just about any development policy – including the Poverty Reduction Strategy Papers (PRSPs) jointly adopted by the IMF and World Bank in 1999 (examined in Chapter 5).³⁹ More generally, both organizations and many donors began to rely more on the active participation of civil society to achieve their development objectives. They were supported in this shift by key state actors, particularly the British and Nordic country directors, whose home governments had embraced the aid effectiveness agenda.⁴⁰ World Bank and IMF staff sought to mobilize new, more active and responsible public and market actors who could pressure their governments for reform, constituting the “demand side” of good governance policy (see Chapter 6).

If NGOs and other critics could charge the IFIs with failure on the grounds of their political interference, the institutions’ staff responded by redefining failure in a different way altogether: studies by Dollar, Svensson and others argued that the failure of programs was linked to political problems in borrowing countries.⁴¹ The adoption of these new political economy ideas, the development of new participatory techniques and the engagement of new civil society actors enabled the IFIs and donors to respond to criticisms of their interventionism by actually expanding their involvement in domestic policy. While this was a paradoxical response, it was an effective one, for it shored up the institutions’ declining authority in several ways. By focusing on the domestic political sources of policy failure, the IFIs deflected responsibility for poor results. IFI staff also had the opportunity to develop expertise in the arena of political economy, and thus to justify their expansion into new terrains. At the same time, by relying more on political techniques such as participation, and (eventually) country ownership, they were able to supplement their expert authority through appeal to popular support within borrowing countries themselves.

The limits of technical universals

As I discussed in Chapter 3, organizations often seek to govern in the name of certain universal values or principles, and to govern through their use of techniques and practices that they deem to be of universal applicability. IMF and World Bank leaders had largely eschewed any overtly moral framing of the organizations’ universalist aspirations in the 1980s, relying on technical economic principles as the basis of their claim to universality. This was an approach that fitted well with their claim to expert authority. Yet this was also a vulnerable strategy precisely because its authority relied so heavily on the promise that one set of economic principles could be applied universally.

The events of the 1990s and early 2000s were seen by many as a major test of these universal economic ideas – a test that the IFIs were widely viewed as having failed. The Asian crisis provides a particularly stark example of the kind of erosion that began to occur in the foundations of economic orthodoxy. In 1993, the World Bank published a report entitled “The East Asian Miracle,” which sought to make sense of the remarkable economic growth in this region.⁴² As Robert Wade has so effectively demonstrated, rather than recognizing the positive role played by activist East Asian states in supporting this success, the report’s authors instead chose to downplay it: “The result is heavily weighted towards the Bank’s established position, and legitimizes the Bank’s continuing advice to low-income countries to follow the ‘market-friendly’ policies apparently vindicated by East Asia’s success.”⁴³ Even in the face of consistent pressure from Japan and significant evidence contradicting the Bank’s position, Bank staff and management held onto their singular, universalist conception of sound economic policy.

The IMF responded to the Asian financial crisis of 1997–8 with the same approach, applying policies that had been used to deal with earlier Latin American crises to a radically different policy environment. They also used the crisis as an example of what can go wrong when economies do not fully embrace the strict free-market model, and sought to re-introduce more Anglo-American economic policies to the region. Yet this time, the universalist model came under enormous strain. The IMF’s policies were blamed for worsening rather than resolving the crisis. In 1998, the World Bank published a report that largely blamed the intensity of the crisis on the IMF and the US Treasury.⁴⁴ There was no longer a consensus in Washington on economic policy. Many of the same economic assumptions about low inflation and economic liberalization continued to underpin IFI policies, yet they were no longer as universally accepted as they once were. The principles that had been so confidently relied on since the early 1980s were now the subject of widespread debate and problematization both within and outside the IFIs.

As the economic universals of the structural adjustment era began to erode, they were not replaced by a dramatically new paradigm. Instead, two different responses to this dilemma emerged: IMF and World Bank staff began to pay more than lip service to the idea that there was a diversity of different economic situations and began to focus more on particular contexts, leading to the strategy of ownership; and they began to redefine universals in more normative and flexible terms – to include norms of good governance and standards of best practice – producing the strategy of standardization. Although IMF staff had always rejected the claim that they had applied a “one size fits all” approach to adjustment problems, after the backlash from the Asian crisis they were forced to modify their approach.⁴⁵ At the World Bank too, beginning in the late 1990s, there was increasing concern with ensuring that policies on good governance, for example, were carefully tailored to specific local needs and concerns.⁴⁶

There had also been a long history of interest in the problem of what Robert McNamara called “political will” – the need for domestic governments to buy in to Bank and Fund programs. By the late 1980s, this had been refined into the concept of “country ownership,”⁴⁷ which the OED first attempted to measure in 1992,⁴⁸ and which the IMF adopted as a key concern in 1998.⁴⁹ The strategy of country ownership was double-edged: it promised more attention to local political concerns in order to attain political buy-in, but it also placed greater responsibility for program success on domestic leaders. The practice of fostering ownership thus allowed staff at both organizations to shift much of the blame for policy failure onto domestic political systems; at the same time, it provided techniques for bringing local political leaders and civil society into the programs as more active and responsible participants.

Another major response of the IFIs to the erosion of their technical universals was to supplement them with a different kind of universal. As programs moved increasingly into the business of rebuilding institutions as well as reforming policy, staff sought to redefine the universal principles of the global economic order to include good governance practices as well as macroeconomic policies. As I will discuss in Chapter 6, these new global standards were different from the previous technical universals in several respects: they were broader in scope, explicitly covering political, social and economic issues; they were justified in moral as well as technical terms; and they took a more flexible and visibly constructed form than the economic principles that they supplemented.

The 1990s thus witnessed both the culmination and the decline of the structural adjustment era’s economic universals. While the problematization of these universals was a powerful blow to IFI authority, the two new strategies that have emerged in response have both succeeded in re-establishing it in several ways. Renewed attention to particular contexts and local ownership has required the creation of a range of new forms of expertise for applying political economy frameworks to understand and act upon local contexts. Moreover, by framing these new universals in moral as well as technical terms, the IMF and World Bank leadership has also sought to create a more robust basis for their global authority.

Debating success and failure

Like most international organizations (IOs), the IFIs’ claims to expert authority also relied on their ability to demonstrate at least a certain measure of “success” in their programs. As far back as McNamara’s time, there had been enormous emphasis on making sure that programs were seen to be successful.⁵⁰ Yet, as I discussed in the previous chapter, there were persistent difficulties with measurement and evaluation, as both IMF and Bank staff struggled with the limits of their abilities to calculate and evaluate their programs. Both organizations had always struggled with a paradox of sorts: they could measure those things that were relatively easy to quantify, such as inputs or narrow objectives, and sacrifice measuring less tangible aspects of their programs (in particular the role of influence); or they could focus on these more slippery factors, but in doing so find themselves struggling with measurement challenges.

Fund staff responded by experimenting with different methodologies, while in the 1990s, the OED at the Bank introduced a more sophisticated metric for measuring success, which included an initial assessment of the riskiness, “demandingness” and complexity of the project, and an assessment of sustainability and institutional development.⁵¹ By the 1990s, as the Bank began to assess the success of those projects and programs initiated in the 1980s and to use more sophisticated metrics to do so, they found their success rates dropping precipitously – from the 80 to 85 per cent range to below 65 per cent in the early 1990s.⁵² The 1992 Wapenhans Report was particularly critical in its assessment of the poor success rates at the World Bank, and intensified the search among staff and management for ways of improving them.⁵³

In the course of the 1990s, discussions of the problem of failure began to grow more prominent at the World Bank and within the aid community. At a popular level, critics from both the left and the right were vigorous in condemning the Bank for what they saw as wholesale failure: NGOs and groups such as “50 Years is Enough” attacked the IMF and World Bank for inflicting untold damage on the global poor through their neoliberal policies. On the right, there was a growing chorus of critics, many in the US, who argued that aid was no longer necessary in a world of integrated capital markets.⁵⁴ In academic and policy circles, a host of studies examined the causes of success or failure in a development project.⁵⁵

The most influential among them included Dollar and Svennson’s “What Explains the Success or Failure of Structural Adjustment Programs?” and the Bank’s own report Assessing Aid: What Works, What Doesn’t and Why, headed up by David Dollar.⁵⁶ These studies adopted different metrics from the ones then being used by the OED – focusing on whether policies created “sound policy environments” defined in both macroeconomic and institutional terms.⁵⁷ While their conclusions differed in some respects, they both raised serious concerns about the low levels of Bank success and focused on domestic political and institutional factors as the key reasons for program failures. These and other studies also questioned the effectiveness of conditionality – particularly structural conditionality, which had been the dominant technique of the structural adjustment era.⁵⁸ They suggested that without local ownership and domestic institutional capacity to implement policies, increasing the number of conditions was at best pointless, and at worst counterproductive. Assessing Aid suggested some significant policy changes, including radically reducing funding to states that did not already possess the right “policy environment.”⁵⁹

These internal critics proposed a different set of criteria for both operationalizing and evaluating aid, using effectiveness as the central metric – a metric that relied heavily on political economy factors. This was a classic example of a hot rather than a cold debate, since the very question of what counted as evidence of success and failure was open to debate – simply getting loans out the door and obtaining a reasonable rate of return was no longer enough to make a project count as a success.⁶⁰

Did these studies both within and outside of the Bank discover an underlying pattern of objective failures in Bank and donor lending? Yes and no. They certainly did point to some troubling findings, but this does not mean that these were the only conclusions that they could have reached: it was partly because staff and scholars started to change the metrics for evaluating success and failure (focusing on institutional development, sustainability, policy environment, etc.) that they began to discover more failures.⁶¹ And it was because of the theoretical lenses that they used in these studies that they diagnosed the problems and solutions as they did.

Both of these studies drew heavily on public choice theory to explain program success or failure, and on new institutionalist economics to propose solutions. From a public choice perspective, borrower governments will generally try to “game” the system by promising reforms that they may not intend to undertake. With aid being fungible (aid dollars allocated for one project freeing up government funds for something else), there are few ways for agencies to control the government’s actions and ensure “success.” Hence, the best way of guaranteeing that the desired outcome is achieved is to lend exclusively to countries that are most likely to use aid effectively – which, these studies suggest, are those that already have “sound” as opposed to “distorted” policy environments. Institutionalist economics, in turn, suggests that sound institutions are also necessary for good policy: hence aid should be directed selectively towards those states that are already in possession of the rule of law, a capable public sector and a low level of corruption.⁶²

These debates on aid effectiveness did not reach as deeply into the IMF. Nonetheless, some of the same ideas that were shaking up the World Bank’s policies, such as ownership and selectivity, also started to take hold at the IMF. In 1997, the IMF embarked on two different reviews of ESAF – the highly conditional and longer-term lending facility that its poorest members relied on. One was an internal review, conducted by the Policy Development and Review Department.⁶³ The other was an external review, which included among its members Paul Collier, a major figure in the aid effectiveness debate who became an influential actor at the World Bank when he started working there a year later.⁶⁴ The two reviews had very different mandates. The internal one provided a very neoclassical analysis of the successes and failures of ESAF programs and recommended budget cuts, inflation fighting and other neoliberal staples. The external review considered the social impact of ESAF programs and, most interestingly, their capacity to foster country ownership, recommending more attention to poverty and social impact and more genuine openness to negotiating with borrower governments.

The external report used different criteria for assessing the institution’s programs’ success or failure – considering its social impact – and asked the IMF to do the same, drawing on the World Bank’s expertise to do so. Both reviews, moreover, found that political factors had a considerable effect on the success or failure of ESAF policies, forcing the institution to reconsider the key determinants of policy viability. Both the IMF and the World Bank thus found themselves having to redesign the metrics through which they judged success and failure; both began to pay more attention to political factors as crucial; and both also sought new measurement and evaluation techniques that could better capture these more complex dynamics. At the World Bank and among many donors, this search brought them eventually to the attempt to measure development results.

The problem of contingency

One of the subtlest but most insidious challenges to the IFIs in the 1990s and 2000s was the problem of contingency. This was not simply because crises in finance and development occurred: crises do happen and can almost always be blamed on exogenous factors. The problem was that the institutions had not factored the possibility of such happenings into their governance strategies: they had been caught napping. Their linear conceptions of policy time did not provide a way of coming to grips with disasters except in the most reactive of ways. Their promises of predictive power – a key part of much economic theory – turned out to be hugely overstated in the face of these unexpected events.⁶⁵ Although in the past, staff might have been able to adapt to unexpected results by further extending the time horizon or creating specialized facilities, the recalibrated measures of success and failure seemed to suggest that something more profound was going on: if the problem was institutional capacity or the political environment in borrowing countries, and not just the narrow economic factors the IMF and World Bank had been focusing on, then these organizations would need to find ways of engaging those more complex issues in an increasingly volatile context. The contested failures in African development, and in Asian and later global finance, thus helped to precipitate a series of more profound problematizations about how to do the work of economic governance in a more contingent environment.

It was in the context of their grim assessments of development success in Africa that the World Bank’s OED staff first started to make systematic use of the idea of risk. In 1996, their annual report was entirely structured around the idea of a risk-based assessment of project success or failure.⁶⁶ Introducing yet another series of new metrics, they sought to categorize all programs in terms of their level of risk and reward, and then map the patterns of risk across regions, sectors and types of programs. Poverty-oriented, institution-building and structural programs were all deemed to be high-risk (but also potentially high reward). The goal of the report, however, was not simply to measure and map such risks, but ultimately to propose ways of reducing them – in order to increase the Bank’s success rate back to 80 per cent. How were they to do so? Here the aid effectiveness literature discussed above became very useful in suggesting that greater selectivity in lending could be the key to reducing the failure rates.

The Bank staff’s perception of the Asian financial crisis and the AIDS crisis in Africa also precipitated a related use of the ideas of risk and vulnerability as a way of conceptualizing contingency: part of what was so shocking about both events was the way in which they not only aggravated existing levels of poverty, but also forced people who had climbed out of poverty back into penury. This again upset any conception of poverty reduction as a linear process. As I will discuss in Chapter 7, staff in the social development unit responded by redefining poverty as risk and vulnerability, an idea that ultimately became a central feature of the 2000–1 WDR.

The Asian crisis also forced the issue of risk onto the IMF’s agenda, having put into question the organization’s capacity to effectively predict and prevent major economic crises. It was in the aftermath of that crisis that the organization introduced its Financial Sector Assessment Program (FSAP), as part of its standards and codes initiative, which was designed to assess a range of different financial risks within participating states and propose ways of mitigating them. Yet, despite much discussion of developing better mechanisms for predicting and preventing future crises, the IMF did not really begin to take seriously the challenges of risk and vulnerability, particularly for low-income countries, until after the 2008 financial crisis. It was only in the aftermath of that contested failure that Fund staff began to focus on the growing impact of external shocks on low-income countries, a problem that they sought to address by assessing their vulnerability.

If financial crises and the failures of African development challenged the capacity of IFIs to govern contingency, then risk and vulnerability assessment and management seemed to promise a new more effective way of governing the vicissitudes of financial and development reform.