Introduction
Benchmarking is a governance instrument that – among other things – establishes a shared reference point among otherwise distinct and distanced actors. The use of benchmarks and indicators to monitor and govern from afar has proliferated over the last two decades, and the international development apparatus is a striking case in point. Since the early 2000s, donor institutions have increasingly relied on benchmarks in their engagement with client countries, partly because of new partnership formations and the shift towards greater emphasis on ownership and post-hoc conditionality.Footnote 1 Benchmarks and indicators serve to monitor and govern by rendering the world legible, transforming, for example, different countries’ complex socioeconomic and political features into numerical rankings.Footnote 2 Quite a lot is at stake in such benchmarking processes – both in producing and responding to them. As the editors note in their introduction to this Special Issue, benchmarks can neutralise or suppress political tensions and universalise particular interests through a quantitative and seemingly objective language.Footnote 3 The editors’ conceptualisation of benchmarks invites us to reflect on the ‘politics of benchmarking’ and to assess how benchmarks work as tools for ‘governing at a distance’.Footnote 4
We follow the editors’ focus on the politics of benchmarking and explore in more detail how benchmarks can be said to be or represent authority in global governance. We do this because extant literature on benchmarks has tended to attribute authority and power to global benchmarks without detailed analysis of what this implies for benchmarked actors. We specify what is meant by the concept of authority and foreground that it first and foremost describes a relationship between a superordinate and a subordinate actor, where the latter defers to the former. It follows that to assess the role and functioning of benchmarks, we have to look at the uses of benchmarks and at the reactions to them by benchmarked actors.Footnote 5 We draw on a case study of how the World Bank uses one of its benchmarks – the Country Policy and Institutional Assessment (CPIA) – in its ‘governing’ and relationship with two African countries, Ethiopia and Malawi.
Ethiopia and Malawi represent two interesting cases as both – but for different reasons – have experienced friction in their partnership relation with the World Bank, thus alerting us to local contestation and agency at the ‘receiving end’ over the global policy prescription embedded in the CPIA. The World Bank and its sister institution, the IMF, have disengaged their operations in both countries in certain periods, and both countries – though more so in Ethiopia – have had overt ideological disagreements with these institutions. The CPIA and the policy prescriptions that flow from it are more likely to be adopted in countries whose governments to a larger degree already share the World Bank’s policy outlook. By focusing on two countries where the governments have resisted World Bank policy, and where the World Bank have discontinued budget support in certain periods, we can get a better grasp of the scope, and limits, of the authority of the CPIA as a benchmark, and of the World Bank as an organisation more generally.Footnote 6
We proceed as follows: We first present a brief discussion of the concept of authority, which prepares the ground for the empirical analysis of the CPIA. We focus on how the World Bank describes its role, including how it is used to allocate funds, to signal incentives to developing countries, and to work as a diagnostic tool for policy improvement. We then present findings from our two case studies of Ethiopia and Malawi. We focus on the role of the CPIA in allocating funds, in signalling incentives, and in shaping the policy dialogue and the Bank’s country-specific strategy. We find that the CPIA is not – and contrary to what one would expect from the CPIA’s nominal function and the literature on benchmarks – a very important tool for signalling incentives and allocating funds, or shaping the policy dialogue or the World Bank’s strategy in these two countries. Rather, the CPIA is used highly selectively as one factor among many in the negotiations between World Bank staff and governments. These findings prompt us to reflect on the use of benchmarks and what type of authority that benchmarks such as the CPIA can be said to represent. We argue that the CPIA is best understood as a governing tool that the World Bank uses to govern and maintain its authority vis-à-vis both aid recipients and donors.Footnote 7 Presenting findings from our own study and that of others, we argue that the CPIA helps establish – together with other tools – the World Bank as an actor that embodies global authority on development, even if the CPIA (and other universalising governing practices) are less important as a tool in the World Bank’s country-specific engagement, as in Ethiopia and Malawi, over the contents of government policies.
Authority and benchmarks
A key claim in extant literature is that benchmarks do, in fact, represent or contain power or authority. Davis et al. argue, for example, that: ‘The burgeoning production and use of indicators in global governance has the potential to alter the forms, the exercise, and perhaps even the distributions of power in certain spheres of global governance.’Footnote 8 Alex Cooley notes that organisations that produce benchmarks and rankings can be ‘judges’ and ‘regulators’ and that benchmarks have the ‘power to informally regulate global institutions and practices …’.Footnote 9 Similarly, Sally Engle Merry argues that indicators – what we here refer to as benchmarks – ‘shift responsibility for governance from those in power to those who are governed’.Footnote 10 This focus is understandable, since most studies have analysed the ‘ranking frenzy’ as a novel macro-phenomenon where a range of actors uses benchmarks as a tool to govern and hierarchically order states.Footnote 11 Nonetheless, detailing the prevalence of benchmarking as a form of governance at a global scale is not the same as demonstrating that benchmarks are authoritative in the sense of inducing deference and shaping the behaviour of benchmarked actors.
It should be recalled that authority designates first and foremost a specific type of relationship between actors. Crucially, most analysts agree that ‘recognition’ is a key component of authority. David Lake notes that authority ‘does not exist without recognition’,Footnote 12 and Deborah Avant, Martha Finnemore, and Susan Sell similarly note that ‘Authority is created by the recognition, even if only tacit or informal, of others.’Footnote 13 This view of authority builds on Richard B. Friedman’s discussion, where authority designates a relationship between a superordinate and a subordinate actor, which is recognised, and where the latter defers to the former.Footnote 14
In this perspective, authority concerns the relationship between the producers/users of benchmarks, on the one hand, and the targets of benchmarks, on the other. Benchmarks are tools that can shape that relationship, but it depends on the context of its use. It also means that authority does not spring magically from some ideal-typical ‘source’ of authority, whether it is expertise, legal rules, or moral standing.Footnote 15 Rather, authority is something that is at stake in relationships where part of the competition revolves around what is to count as a ‘mark’ of authority.Footnote 16 While we think it makes sense to foreground the positions and resources of the actors involved in producing and using benchmarks (rather than analysing whether benchmarks themselves as having authority), we have to analyse how actors can use benchmarks to construe or strengthen a position of authority. This is not to say, however, that benchmarks can never be authoritative, only that to explain their authority, we must move beyond assertions about benchmarks’ power to universalise and to substitute numerical descriptions for complex explanations, and explore the positions and power resources of those actors that go into establishing and using these benchmarks. In other words, what needs to be accounted for is not merely how a benchmark is made authoritative as a description or definition of some phenomenon, but also how that benchmark is used by different actors to try to shape the behaviour of their intended targets, and how intended targets of benchmarks react to these.
This has important implications for our empirical analysis, as we focus on how the actors involved use and respond to the CPIA in Ethiopia and Malawi. We also seek to be sensitive to the possibility that an actor that may recognise the epistemic authority of the World Bank, and of the CPIA as a benchmark, and yet decide to act in ways that go against such authority because of domestic political pressure or other considerations. Likewise, there is also the possibility that governments act in accordance with the benchmarks not because of the power of benchmarks but because the governments agree with the benchmarks’ policy prescriptions.
But before we move on to analyse the uses of and reactions to the CPIA, we need first to present the CPIA and the rationale for its design and use as seen by the World Bank.
The formal order of the CPIA
The CPIA consists of 16 criteria grouped into four equally weighted clusters.Footnote 17 It is designed to measure a country’s overall policies and institutional quality. It measures only those aspects that are under a country’s control, and it focuses on actions and performance, not plans and intentions. By itself, the CPIA is merely one among many instruments the World Bank uses to assess partner countries. What makes the CPIA important is that it has always been tied to the allocation practices of the International Development Association’s (IDA), which is the Bank’s soft loan branch offering concessional lending on terms preferable to other types of loans offered by the Bank.Footnote 18 The CPIA provides the key input to the Bank’s Performance Based Allocation (PBA) mechanism, where allocation of funds is based on ‘needs’ (measured by population size and poverty levels) and ‘performance’ (measured by the CPIA). The CPIA and the PBA together transforms a political question – how to allocate scarce resources from donors to recipients via the Bank – into an allegedly rules-based practice, based on the content and quality of countries’ policies and institutions.
The CPIA matters for more than the allocations, however. It also signals incentives – partly via the PBA system – to client countries to adopt policies that conform to the policy prescriptions contained in the CPIA, to anchor the World Bank’s ‘Country Assistance Strategy’, and to underwrite the World Bank’s policy dialogue with governments of client countries.Footnote 19 The structural setup of the CPIA reflects the constitutive elements of benchmarking practices as identified by the editors of this Special Issue: reification by converting the diverse and subjective into legible, objective properties; commensuration by aligning the different units according to a common metric; symbolic judgement by ranking complex social, political and economic phenomena.Footnote 20 But it also goes beyond these in that it is directly linked to economic incentives, where higher CPIA scores translate into more funding through the PBA system.Footnote 21
The CPIA consists of a set of criteria that represent the different policy and institutional dimensions that the World Bank deems important for effective poverty reduction and economic growth. The criteria have grown in number and scope, reflecting the institution’s evolving priorities and policies. From an initial focus on financial management and economic indicators, it was expanded and made more comprehensive during the course of the 1990s. In 1997, governance was explicitly introduced through a new criterion on legal and regulatory framework. Here, concern with corruption was central, as evidenced by the reference to the importance of paying attention ‘to the implications for corruption of public sector administration’ in assessing all elements – tax reforms, volume and composition of expenditures, public expenditure management, and civil administration’.Footnote 22 The following year, the CPIA was again expanded and refined as it now came to reflect ‘an agreed set of Bank wide criteria for promoting growth and poverty reduction’.Footnote 23
The World Bank started to disclose its ratings in the early 2000s. The reasons for disclosure are manifold, but for our purposes, two of them are particularly important. The first has to do with the World Bank’s ambition to broaden the role and scope of the CPIA and improve its incentive structures. In order for the CPIA to creative incentives for compliance and have bearing on policymaking processes, the CPIA ratings have to be known. The World Bank has noted, for example, that it is important to simplify how the assessment in the CPIA factors into the PBA in order to ‘make the system more effective at the country level’.Footnote 24 This is part of the reason for the Bank’s decision to make the CPIA ratings public. As one review notes about the CPIA and PBA: ‘Transparency and clarity are, of course, extremely important to the rationale for using a performance-based formula at all. The easier the formula is to understand, the more clearly can its incentive aspects be communicated to the eligible countries.’Footnote 25 The second has to do with the introduction of the Poverty Reduction Strategy (PRS) as a new mode of engaging with client countries, giving them more ‘ownership’ of policy formulation and implementation. This meant abandoning the Structural Adjustment Programmes (SAP) through which the World Bank imposed its readymade policies of privatisation and liberalisation onto multiple and diverse counterparts.Footnote 26 This approach, as part of the so-called Washington Consensus,Footnote 27 was highly criticised for being top-heavy and providing blueprint, one-size-fits-all models. The PRS-system, rolled out from 1999, set out to be more inclusive and context based, as recipient states now were to devise their own, national development strategy, based on principles of ownership and participation, to provide the foundation for the World Bank’s financial support. In this context, the World Bank describes the role of the CPIA as a ‘diagnostic tool’ that ‘indicate[s] areas where attention needs to be focused and thereby influence the focus of the IDA-supported program’.Footnote 28 Disclosing the CPIA, and connecting it to the incentive mechanism of performance-based allocations while widening its scope to signal a wider policy focus, can thus be seen as an instrument for the World Bank of combining greater freedoms and responsibilities in planning and implementation for client countries while at the same time securing an access point for governing by the World Bank.Footnote 29
The production of the CPIA is done every year for each IDA eligible country by World Bank staff, that is, the Bank’s respective country teams in cooperation with the region’s Chief Economist. This is of some importance, as the actors engaged in assessing a country’s performance are the same, at least at the country level, as those using the analysis when putting together the lending portfolio. The World Bank shares with the government the so-called ‘write-ups’ that underpins and justifies the numerical score given on each of the criteria contained in the CPIA. The government, however, is not involved in negotiating the benchmarking exercise and the World Bank’s analysis.
The CPIA in practice: Incentives and policy dialogue in Ethiopia and Malawi
The World Bank presents the CPIA as performing two core functions: (1) allocate funds and signal incentives to partner countries for improved performance; and (2) shape the country-level policy dialogue, and the formulation of the World Bank’s programmes in particular countries through the Country Assistance Strategy (CAS).Footnote 30 The fact that the World Bank identifies these functions as integral to the CPIA is an expression of the importance that the World Bank attaches to it in its management of relations with recipient governments. We focus on both allocations and incentives, and on the policy dialogue and the CAS because they touch on different aspects of the relations between the World Bank and the governments of Ethiopia and Malawi. The first centres on allocations and the signalling of incentives, where we would expect the CPIA to be a powerful tool to shape behaviour. The second centres on the CPIA’s role in shaping the policy dialogue between the World Bank and in formulating the CAS, where we would expect the CPIA to play a different, and less direct, role in shaping the behaviour of the government than in signalling incentives and determining aid allocations.
Context: Ethiopia and Malawi
Ethiopia’s relations with international donors date back to the 1950s. It was the first African country to receive World Bank assistance, in 1950. Today it tops the worldwide list of countries receiving aid from the US, UK, and the World Bank, receiving $3.5 billion on average in recent years, which represents 50 to 60 per cent of its national budget. As Ethiopia never was colonised, it had a state apparatus when it got engaged with donors. Thus it was quite unlike most other aid-recipient countries that only started to receive aid upon independence. One implication is that the Ethiopian authorities in past and present have seen aid relations as the meeting of equals, with a government more assured of its own policy directions, its entitlements to the set development agenda and of its stature vis-à-vis the international donor community – as compared to most other aid recipient countries.Footnote 31
Ethiopia has made development gains over the last decade in the health, education, and food security sector, and the national economy has seen a consecutive double-digit growth over the same period. Yet, Ethiopia still remains one of the poorest countries in the world: roughly 30 per cent live below the poverty line of $1.25 a day, and many are vulnerable to food insecurity. Over 80 per cent is dependent on subsistence agriculture and its fast growing population and climate change put greater pressure on natural resources and land, being the cornerstone of the country’s growth. Ethiopia is seen to be a poor performer on governance issues, both according to the CPIA and other donor assessments. Yet, the same assessments deem Ethiopia to do well in terms of economic growth and service delivery. International donors agree, for example, that Ethiopia’s good achievements are due to it having one of the most ‘pro poor budgets’ in Africa.Footnote 32 Ethiopia is therefore interesting inasmuch as it has established policies that go against the basic tenet of the CPIA, and indeed international development discourse in general, yet have succeeded in generating economic growth.
Malawi is one of the world’s most densely populated and poorest countries. The population is at 13.6 million people, with around 80 per cent of the population living in rural areas.Footnote 33 By 2010 – the period where our empirical analysis in both countries was concluded – donors contributed around 40 per cent to Malawi’s annual budget. In the period from 2004 to 2009, Malawi saw substantial economic growth, averaging 7.5 per cent. This economic growth period was preceded by a period of economic mismanagement where the IMF defined Malawi ‘off-track’, causing both the IMF and other donors, including the World Bank, to either suspend or drastically reduce their development assistance. Partly as a consequence of this reduction in development assistance, Malawi borrowed heavily in the private market, accumulating high levels of debt. In 2005, Malawi completed the IMF ‘staff monitored program’ and reached the completion point of the HIPC initiative in 2006, according it debt relief amounting to over US $2 billion.Footnote 34
Corruption and financial mismanagement has been a key source of friction between the government and the World Bank (and other donors). One study argues that that mismanagement of public resources is so widespread that the negotiation over the government’s budget is like a ‘theatre’: ‘All the actors, from civil society, government, and donors seem aware that many of their statements and actions have little bearing on actual distribution of resources. Yet, all stakeholders ‘act’ as if the budget planning and formulation will actually have a bearing on the actual implementation and distribution of resources.’Footnote 35 As in Ethiopia, the government has clashed with the World Bank on major policy issues: in 2005, following years of drought and famine, President Mutharika introduced a programme for subsidies of fertilisers to boost agricultural production against the recommendation of the World Bank and other donors.
Allocations and incentives
From 2005 to 2008, Ethiopia’s annual, average CPIA score was 3.4, on a scale from 1 to 6, which is quite low (states with an average below 3.2 are categorised as fragile).Footnote 36 In 2005, moreover, the World Bank and other donors were highly critical of the government’s handling of the post-2005 election turmoil, whereupon several donors reduced their support. This political situation caused the World Bank to terminate its budget support mechanism, which is difficult to interpret as anything else than the World Bank signalling discontent with the government’s handling of the situation. Nonetheless, we see no cuts in the World Bank’s overall allocations to Ethiopia in this period. Rather, IDA funding increased from US $449.9 million to 634.5 million over the same period.Footnote 37 This increase in allocation occurred at the same time as the World Bank moved from direct budget support into more direct programme support. Since IDA allocations are determined annually, with the CPIA score as the main determinant, this is a strong indication that the CPIA does not determine aid allocations in a way that clearly signals incentives to behave in keeping with the prescriptions of the CPIA. Part of the explanation for this decoupling between performance assessment on the one hand, and actual allocations, on the other, is that other factors kick in and shape the allocation to each country, including changes in ‘performance relative to other countries, changes in eligibility for IDA resources and for IDA grants, and availability of IDA resources’.Footnote 38 We mention this in some detail because it raises the issue, discussed in more detail below, that we should be attentive to how the World Bank presents itself and its operations with a view to securing its credibility and authority vis-à-vis both recipients and donors at the same time.
From 2005 to 2008, there was a significant decrease in the CPIA score on macroeconomic management, from 3.5 in 2005 to 2.5 in 2008. According to the World Bank, Ethiopia’s growth is challenged by high domestic inflation and a difficult balance of payment situation with weak foreign exchange reserves driving fuel and food prices up. In order to mitigate and control the high domestic inflation rates, the government has removed fuel price subsidies, and the National Bank has raised the minimum reserve requirement on commercial bank deposits. Nevertheless, the World Bank asserts that the macroeconomic situation remains tenuous,Footnote 39 and the decrease in CPIA score on economic management is made with reference to an incomplete sequencing and implementation of what the Bank sees as a sound policy response by the government.Footnote 40 And yet, the Ethiopian economy experienced strong and broad based growth over the last decade, averaging 10.9 per year in 2003 to 2012 compared to the regional average of about 5 per cent, producing positive trends in reducing poverty in urban and rural areas. Among critics of the World Bank’s use of the CPIA to determine aid allocations, this is a key point: Many countries, including Ethiopia, have been able to generate stable economic growth and poverty reduction without conforming to the prescriptions contained in the CPIA.Footnote 41 All interviewees agreed that these measures were undertaken on the government’s initiative to improve the macroeconomic situation, and not – by their own admission – as an attempt to improve the CPIA rating. Both government and donor representatives also report that the government would never alter its policies just for the sake of accommodating donors or to increase its CPIA to gain more IDA funding. Instead, interviewees held that the government would rather opt to lose development assistance than accommodating donors’ demands and policy prescriptions.Footnote 42
A similar pattern can be found in Malawi, where CPIA ratings have been quite stable over time, with some significant changes on the cluster on Economic Management, where fiscal policy improved from 3 to 3.5 between 2006 and 2007, and macroeconomic management also improved from 3 to 3.5 from 2005 to 2006. IDA allocations, however, have changed quite significantly in a negative direction in the same period, from US $116 million in 2005 to 47.2 million for 2006. The CPIA ratings cannot explain such large changes from one year to another. Ceteris paribus, Malawi would receive a higher allocation with an improvement in CPIA rating. As with Ethiopia, both primary and secondary sources suggests that even in the case of a strong link between performance and allocations, it is unlikely that it would contribute to Malawi changing its policies or institutions in a certain direction.Footnote 43 A case in point is the decision to introduce subsidies of fertilisers against the explicit recommendation of the World Bank. As in Ethiopia, the government appeared to have made headway in economic terms by going against the policy prescriptions of the World Bank. As reported by the New York Times:
Over the past 20 years, the World Bank and some rich nations Malawi depend on for aid have periodically pressed this small, landlocked country to adhere to free market policies and cut back or eliminate fertilizer subsidies, even as the United States and Europe extensively subsidized their own farmers. But after the 2005 harvest, the worst in a decade, Bingu wa Mutharika, Malawi’s newlyelected president, decided to follow what the West practiced, not what it preached.Footnote 44
Interestingly, the government’s decision to subsidise fertilisers did yield tangible results in corn production, making Malawi a net exporter of corn rather than having to rely on food aid from donors in a matter of years. In response to this decision, the World Bank reduced its support from US $40 to 30 million to signal discontent with the policy, in coordination with other major donors. This did not entail a large overall reduction in aid, however, as it was reportedly a coordinated move: ‘When the Bank is not happy about something, it signals this by reducing its support somewhat, but it allows the other donors to continue its support. It’s a bad cop, good cop thing.’Footnote 45
The lack of a strong or direct link between the CPIA and aid allocations, and also between the policy prescriptions contained in the CPIA and government policy, suggest that the CPIA is not authoritative. If the CPIA had been authoritative, we would expect the governments in Ethiopia and Malawi to ‘defer’ to its policy prescriptions, and the economic incentives associated with the CPIA – via the PBA – would have weighed heavily on the governments’ decision-making. While we should be cautious about over stating our case and note that the CPIA may well induce deference in other countries, our findings are not unique: Other studies have documented that IDA allocations may change significantly from year to year, due to political pressure from key donors in the Bank. Andersen et al. note, for example, that the World Bank decided to withdraw lending to Allende’s Chile in 1970–3; and during the Iran-Iraqi war, the IDA stopped funding to both countries. Moreover, World Bank assistance to Pakistan tripled from US $226 million in 2001 to $860 million in 2002.Footnote 46 Let us now turn to the less direct yet potentially equally important role of the CPIA as a tool that frames the policy dialogue and informs the World Bank’s country-specific operations via the Country Assistance Strategy.
Policy dialogue and country strategy
The CPIA aspires to structure the contents of the policy dialogue at the country level. In Ethiopia in April 2009, at a high-level meeting between government ministers and donor representatives from the World Bank and the IMF as well as bilateral donors,Footnote 47 the macroeconomic situation and the newly-ratified civil society organisations (CSO) law were up for discussion. While creating concerns among donors and squarely being part of the CPIA’s governance benchmarks, the World Bank and the IMF did not comment on the political sides of the CSO law in this High Level Forum. Rather, they commented on the macroeconomic issues where the IMF representative focused on topics falling under CPIA’s Economic Management cluster, and the World Bank representative attended to issues dealing with structural policies, integral to cluster B of the CPIA.Footnote 48
The World Bank representative addressed the structural policies, being central to the CPIA benchmarks, in making a call for stabilising the economy by modernising the financial sector to facilitate private sector activities. This would involve opening up and privatising key institutions such as the banking system and the telecommunication sector, and open up for private ownership of land. This is a case of World Bank staff using the CPIA as a benchmark, where the contents of one government’s policy is held up against that of other countries and the ‘best practice’ implied by the CPIA. At the meeting, however, the government clearly signalled discontent with the World Bank’s policy prescription, and clearly stated that it did not agree: it did not want to liberalise the financial sector, open the banking system to foreign investment, privatise telecommunication, or privatise land ownership. Indeed, the Ethiopian government challenged the contents of the CPIA for reasons of ideology, its alleged subjectivity, and its lack of transparency – seeing it as an ideological package promulgated by the World Bank.Footnote 49 The government sees the debate over structural policies as a form of imposition directly challenging its own ideological orientation.Footnote 50 For example, the government’s celebrated ‘pro-poor budget’ is, amongst others, based on revenues derived in part from its ownership of telecommunication, land, and the banking system – sectors the World Bank would like to see privatised. Consequently, the CPIA is criticised for rewarding policies seen neither appropriate nor desirable by the Ethiopian government. As in Malawi (see below), the government was willing to discuss macroeconomic management, which it itself deems relevant as it directly impacts in creditworthiness, but not governance issues relating to privatisation, which it deems to be ideological and political.Footnote 51
The contestation over ideology relates to the critique over CPIA’s focus on performance, and indicates that the CPIA is not authoritative and hence does not induce deference from the Ethiopian government. The World Bank, for its part, seems to go out of its way to position itself so as to have some say on and role in shaping government policy. For example, the World Bank has called for the resumption of regular budget support even though the political reasons for terminating it is still in place, because the World Bank considers the annual policy negotiations they allow for as far more influential than what is provided for by the CPIA. This suggests that the World Bank is seeking to leverage its position to shape governmental policy, using the CPIA as a tool to manoeuvre and establish a position from where to engage the government on how it should govern. As such, the CPIA is one tool among many that the World Bank uses in its continual negotiation with and efforts to shape the behaviour of, the Ethiopian government.
In Malawi, the case is somewhat different, since there is not the same type of explicit anti-Bank sentiment or ideology within the government. Interestingly, however, the CPIA is known to only a handful of people inside the Ministry of Finance. The responsible section within the Finance Ministry (Debt and Aid) in Malawi has tried to broaden the understanding and implications of the CPIA and the PBA across ministries, reportedly telling colleagues in other ministries that: ‘There is no room for complacency’ as the CPIA ratings affect aid allocations from IDA.Footnote 52 Nevertheless, only those that deal directly with Bank officials and that coordinate with donors are aware of what the CPIA is and why it matters. Because Malawi received budget support in this period, in contrast to Ethiopia, the central arena for policy dialogue was the consultations between the government and the donor group, which includes the World Bank and bilateral donors. This donor group has established a Performance Assessment Framework with detailed targets, actions, and indicators of progress. While the indicators and targets broadly reflect the type of development thinking that underwrites the CPIA, the CPIA as such does not play any significant role in ongoing debates between donors and the government.
The CPIA therefore seems to be central to the discussions between Bank officials and a small group of officials inside the Ministry of Finance, but it is not central to the policy dialogue in broader terms. Bank officials also acknowledged this, noting that: ‘Ideally, government should use the CPIA but they don’t.’Footnote 53 This view was also communicated by officials in the Ministry of Finance, with one informant noting that: ‘We [Debt and Aid in the Ministry of Finance] have tried to sell CPIA to other parts of the government, saying that it matters, but with limited effect.’Footnote 54 The lack of ‘traction’ and even information about the CPIA in Malawi are also reflected in the World Bank’s country specific strategy. World Bank support to Malawi was distributed with 29 per cent for human development, 34 per cent for infrastructure, 24 per cent for agriculture, rural development and natural resources, and so on.Footnote 55 This prioritisation dovetails with the government’s priorities, but it does not harmonise with the CPIA ratings, indicating that the Bank aligns itself with the government rather than with the contents of its own assessment. Or rather, that the World Bank balances the policy implications from the CPIA assessment against the government’s priorities.
Thus far, we have focused on the role of benchmarks in two particular settings, and have assessed whether the CPIA can be said to help the World Bank induce deference. While we see indications that the World Bank is authoritative as regards macroeconomic issues, it is not on other aspects, including liberalisation of key sectors of the economy (Ethiopia), and the use of subsidies (Malawi). Moreover, the CPIA does not appear to be a central tool in the World Bank’s policy dialogue with the government, nor in the formulation of the World Bank’s country-specific priorities. The picture that emerges is one where the World Bank uses the CPIA as one tool among many in its continual negotiation with, and efforts to shape the policies of, these governments. What to make of this? Should we conclude that the CPIA, as a benchmark, have no noticeable effect in these settings, and that the World Bank is not authoritative? Not necessarily. Other governments may relate to the CPIA in a way that is very different from Ethiopia and Malawi. As we argue below, however, the key attributes of benchmarks make them useful for actors that claim a particular type of global or universal authority, but by virtue of this claim to universal application, based on a particular view of how societies should be governed, these benchmarks are simultaneously vulnerable to contestation by actors that have concrete authority in particular settings. Again, we turn our attention to benchmarks as a governing tool, but we now broaden our focus and include in our discussion what the CPIA do to the relationship between the World Bank and donors.
What benchmarks do: Universal vs concrete authority
Benchmarks tend, as the editors of this Special Issue note, to cloak the prescriptive elements that they contain via numerical expressions and can present as description what is prescription. More specifically, benchmarks may constitute what Michel Callon and Bruno Latour call an ‘inscription’ – the translation of particular interests into ‘material form’, or more generally making that which is contested into something uncontested and immutable that can circulate and gain the appearance of objective and fixed form.Footnote 56 Benchmarks help in making some ideas more or less uncontested and immutable, which circulate and gain the appearance of objectivity. Through processes of reification, commensuration, and symbolic judgment, benchmarks give an appearance of facticity to ideas and enable their use by a broad range of actors as they engage in global governance.Footnote 57
The CPIA is a case in point. The core causal claim behind the substantive contents of the CPIA is that a good policy environment is a necessary condition for the effective use of aid. This idea came to prominence with the highly influential study by two World Bank researchers on ‘Aid, Policies, and Growth’.Footnote 58 The CPIA helps produce an image of this causal relationship as objective and uncontested not only by making it numerical and creating a comparable assessment of the quality of each country’s policy environment. It also renders possible a broad range of other activities, such as making allocations according to a rules-based system (the PBA), and creating an infrastructure of seemingly robust data about whether developing countries’ policies are improving that donors and others use in debating the contents of World Bank policy.Footnote 59 The continual practical use of the CPIA by the World Bank and others adds facticity to the data that it contains, even if the causal claim that it builds on is contested in the scholarly literature. With Alain Desrosieres we can say that the CPIA ‘furnish[es] something to lean on, or at least a common language of reference, around which the famous “democratic debate” can be organized’.Footnote 60
The CPIA fills such a function within the World Bank as the indicators and rankings it produces – even when they are disputed and negotiated – serve as the common framework where it is possible to agree and disagree about how to guide, assist, and help developing countries.Footnote 61 For example, there is considerable debate within IDA about the weight of ‘performance’ relative to ‘needs’ in allocating funds. This debate is driven by distinct political views of what constitutes ‘development’ and how it should be best promoted. The PBA system, made possible by the CPIA, here defines the framework within which every actor has to formulate their position. Rather than debating, say, whether country x is to receive more funds than country y, the debate operates at the level of the relative weight of needs to performance in the PBA formula, with the CPIA being a largely unproblematised input to this formula.
Note that the CPIA is here important first and foremost in how the World Bank manages its relations with donors. The World Bank is dependent on funds from donor governments, and the forerunner to the CPIA, established in the late 1970s, focused mainly on recipient governments’ capacity to manage funds: It was established as a governing tool to address donors’ fiduciary concerns over how the funds they contributed to IDA was ultimately managed and spent at the country level. As such, the CPIA is as much about ‘governing’ relations with donors as with recipients as it addresses donors’ concern with performance and fears of misuse of funds and corruption. Moreover, the CPIA enables the World Bank to bring the whole gamut of recipient countries’ domestic policies into its purview without seemingly coming into conflict with its ostensibly apolitical mandate.Footnote 62 It does so by transforming the complex sociopolitical and economic structures of recipient countries into numbers on a total of sixteen indicators, thereby presenting as description what is prescription, and as eminently governable by the tools at the hands of the Bank. The CPIA is in this sense used by the World Bank to (re)produce an image of the Bank as an actor capable of governing and that donors can entrust with their funds in an effort to guide recipient governments towards ‘development’.
This role of the CPIA is distinct from the role it plays in particular borrowing countries. As we saw above, the CPIA was used by the World Bank to try to induce deference in the governments of Ethiopia and Malawi to follow its policy prescriptions on specific policies. The role of the CPIA in managing relations with donor governments is of a different type, having to do with the establishment of a credible image of the World Bank as an international organisation that run on rules and expertise, rather than on the power and political preferences on some key states. Here, the CPIA helps in establishing a different type of authority, one that operates in a universal register, where some rules and norms are held to be good or effective or necessary everywhere, regardless of the particular context.Footnote 63 This is of crucial importance: International organisations and non-governmental organisations (both for-profit and non-profit) establish benchmarks that are typically global in scope, targeting governments, firms, and other actors. Thus understood, benchmarks can perhaps best be seen as institutionalising a universalist language, where some categories and indicators are effective in shaping global policy discourse and presenting as public interests that which is particular interests.Footnote 64 Benchmarks here construct a particular vision of the world and of the objects to be governed, enabling ‘governing at a distance’.Footnote 65
But precisely because the CPIA operates in such a universalist register, used by the World Bank to establish what amounts to an Archimedean point above any specific location or particular context,Footnote 66 it is not tailored to grasp what may produce economic growth in any particular setting. Granted, some benchmarks may be authoritative in particular settings because they are universal. Human rights norms may, for example, be used in particular settings to effect policy change with reference to a universal norm. Their authority to effectuate policy change would nonetheless depend on the configuration of political forces in the context where they are to be applied.Footnote 67 The CPIA, and other benchmarks that implicitly make causal claims about what works, would have to attain the quality of a universal and law-like mechanism in order to overcome the possible tension between the universal and the particular: the laws of physics do apply everywhere, thereby making every particular instantiation a confirmation of the universal. But the CPIA builds on one among many different and competing theories or claims about the drivers of development, which makes its claim to universality vulnerable in the face of evidence that does not fit. As former World Bank researcher Ravi Kanbur has noted, the CPIA measures policies, not development outcomes, and so it cannot account for countries that have poor policies according to the CPIA standard, yet manages to produce high growth rates.Footnote 68 We saw this above in the analysis of Ethiopia and Malawi, both of which have had a level of success in generating growth and poverty reduction (Ethiopia) or boosting agricultural production to become a net exporter (Malawi) through policies deemed inappropriate in the CPIA. This helps explain why the CPIA is not necessarily a significant input factor when a national government makes policy decisions, as such decisions are the outcome of national political negotiations to which external actors are often not privy. As Ann Swidler has noted, ‘Despite donors’ prestige and financial heft, they have more difficulty penetrating and altering local patterns of governance than one might expect. … [W]hat donor organizations offer is received (or seized) within a different social organization, where intended and actual effects differ.’Footnote 69 What makes the CPIA effective in establishing the World Bank with a level of global authority on development, then, hails in no small part from reading out variation between countries by setting up a standard for behaviour that invariably reflect a particular view of society.Footnote 70 When such a standard is used to discuss a general vision for governance and to set terms for resource allocation, it does not interfere with any particular country’s substantive policies on, say, privatisation, social policy, or taxation. But when it is applied in particular contexts and used as a tool to shape governments’ policy decisions, it potentially does come into conflict with domestic political preferences, thereby undermining its efficacy as tool to shape particular countries’ policies.
One way to understand this is in terms of Stephen Hopgood’s discussion of the tension between claiming authority with reference to non-contingent principles applicable everywhere, advanced and defined by cosmopolitans that work (for example) for international organisations, and the ‘claims to national and ethnic loyalty, to historical cultural traditions’. The former is thin, associated with the ‘stranger’, the latter is thick and linked to belonging and substantive identities.Footnote 71 There is therefore quite a lot at stake in how benchmarks are produced and used by organisations such as the World Bank. The CPIA arguably augments the generalised or universal authority of the World Bank, which is important in its relationship with donors. By the same token, however, the CPIA – and benchmarks more generally – help establish a type of relationship with the object to govern that is ‘thin’ and therefore less able to understand and act on the nitty gritty elements in particular contexts.
We should not downplay what this amounts to, since it means that as the CPIA and other universalising tools become the key register for debates about development, and the concrete settings are not considered sources of meaning and policy guidance in their own right. Rather, just as with indexes and benchmarks of fragile states, they impose categories that capture what is lacking rather than what is already there.Footnote 72 International development professionals may represent different actors, be they states, international organisations, non-governmental organisations, business or more loosely organised groups. They nonetheless share a universalised language of indicators and benchmarks, and have professional stakes in being recognised as competent participants in debates about development.Footnote 73 But this is quite distinct from the task of transforming, ‘local patterns of governance’.Footnote 74 We are now in a position to better understand the role of the CPIA. In Ethiopia and Malawi, the CPIA is officially seen by the World Bank as a central policy tool. And yet, in both cases, other factors – whether it is long-established ideological grounds (Ethiopia) or disagreement on specific policy choices (Malawi) – take precedence over the policy prescriptions that are grounded in claims to universality: In both countries, political leaders directly challenged (Ethiopia) or sought to bypass (Malawi) the policy prescriptions of the CPIA. In Ethiopia, reduced CPIA scores and discontinuation of budget support occurred at the same time as an increase in IDA allocations. This apparent paradox can be understood if we consider that the World Bank is using the CPIA to manage relations with donors and to stake a claim to universal authority on development issues. At the same time as the World Bank is seeking to secure this position of a rules-based organisation, which delivers the same policy advice to all aid recipients, however, it also has to stay engaged and seek to have some level of influence in particular settings, which calls for deviation from one-size-fits-all prescriptions. By reducing CPIA scores and moving away from budget support while at the same time increasing IDA allocations, the World Bank manages its relations with donors and recipients. The same dynamics is on display in Malawi, where the World Bank coordinated with other donors to send a signal to the government that they were critical of government policy, thereby cutting support, yet maintained working relations to seek influence on other policy areas.Footnote 75 Indeed, staying engaged and maintaining the partnership relation seems to trump the role of incentives and performance that the World Bank use to describe its operations publicly. The fact that a donor needs a recipient just as a recipient needs a donor illustrates what is often downplayed – that the World Bank, as any bank, is dependent on upholding its lending portfolio and market shares vis-à-vis other donors. The donor and client parties’ joint reluctance to frustrate the overall partnership gives both parts some leverage on the sly of benchmarks’ nominal order.Footnote 76
Conclusion: the limits of global authority
Benchmarks engender a particular type of authority, which is universal in scope, but for this reason it is not necessarily effective in concrete settings. As such, benchmarks are examples of the limits of global authority, and of a mismatch between the register used to describe and explain global challenges such as poverty, on the one hand, and the concrete strategies that may be effective in shaping behaviour in the particular contexts, on the other. We should not, however, conclude from this that benchmarks are not important in particular contexts such as the one under analysis here. As we saw above, the CPIA helps establish the parameters within which policies are formulated and justified as recipient governments engage the World Bank and other donors: In both Ethiopia and Malawi, the World Bank criticised the government’s managing of its economy, and the basis for that critique was taken from the standard defined by the CPIA. Our point is that the CPIA – because it operates in a universal register – is important in establishing the basis of judgment and criticism, but because of this, it is not a very relevant tool for the concrete efforts to shape behaviour or guide policy, where a range of contextual factors (political and economic) tend to dominate.
Against this backdrop, we may ask what benchmarks tell us about the actors that produce them, and whether benchmarks are not part and parcel of an ‘audit culture’ where a particular language or discourse of performance, effectiveness, and accountability is established.Footnote 77 Benchmarks fit what John Meyer and his colleagues call ‘rationalized myths’ – scripts that are enacted to confirm actors’ status as ‘modern’, ‘rational’, ‘accountable’, and so on.Footnote 78 Such a conceptualisation of benchmarks is pertinent, we believe, because it helps explain their prevalence as a technology of governing at the level of global policy discourse even when it appears of little use in concrete settings to address the problems it is purportedly designed to address. Performing or engaging in a particular practice – such as benchmarking – helps produce the image of an actor as competent and authority qua participant in global governance.Footnote 79 A similar view has been advanced among students of development practice. For David Mosse, for example, policy (or benchmarks) should not be seen as coming before and shaping practice – as if an actor’s identity is a given and on the basis of which policy decisions are made and practices subsequently emerge.Footnote 80 Rather, Mosse invites us to see how ‘actors in development devote their energies to maintaining coherent representations regardless of events’.Footnote 81 The CPIA is hailed as important, and defines the infrastructure for the World Bank’s management of relations with donors. It is also used to allocate funds, even if the ‘rule’ of performance based allocations is not always followed. At the country level, the CPIA is said to signal incentives and to inform the policy dialogue, but in practice it is one policy tool among many that the World Bank deploys in its efforts to position itself to have a dialogue with the government to try to shape policy.
