The Janus Face of Valuation: Global Performance Indicators as Powerful and Criticized Public Measures

The article conceptualizes global performance indicators as public measures that are powerful but also receive a wide range of criticism. Global performance indicators derive their power from combining three analytically distinct elements: (a) commensuration (comparing performances on a common metric), (b) visual simplification (presenting performances in an appealing format), and (c) serialization (framing performance as a continuous developing property). However, the very same elements are often subject to criticism. The producers of global performance indicators, therefore, defend method‐ ologies and the validity of commensurated numbers, meet charges of visual oversimplification by professing sobriety and nuance, and balance temporal continuity and discontinuity. By conceptualizing global performance indicators as powerful and criticized public measures, the article draws attention to the Janus face of valuation, which the producers must tackle continuously. Investing considerable time, energy, and resources, these organizations are a key feature of today’s vast indi‐ cator culture.


Introduction
Global performance indicators (GPIs) have proliferated at an astonishing rate over the past decades and are shaping the modern world, which seems beholden to an all-encompassing "indicator culture" (Merry, 2016, p. 9). Here are just a few examples of the vast number and great diversity of GPIs published nowadays (Lokot & Wijermars, 2023;Mennicken & Espeland, 2019): the influential Trafficking in Persons Report by the US State Department rates countries' efforts at combatting human trafficking (Kelley & Simmons, 2015); the Aid Transparency Index by the NGO Publish What You Fund has arguably moved the field of foreign aid by evaluating the commitment of donor organizations to being open about their funding activities (Honig & Weaver, 2019); the OECD's much-discussed Programme for International Student Assessment (PISA) monitors the degree to which national education systems are dedicated to "quality education" (Landahl, 2020).
The ubiquity of GPIs can lure us into simply taking the indicator culture for granted. Social science research certainly has shown a tendency to treat GPIs as social facts, albeit regrettable ones, and to contextualize their proliferation in larger transformations, which are used as structural explanations. Two larger transformations stand out in current debates. First, many work backward from a widespread "trust in numbers" (Porter, 1995) that apparently renders quantitative judgment objective. Second, the proliferation of GPIs is often discussed with regard to unequal "international conditions" (Gutterman, 2014, p. 392), particularly in "political and economic contexts" (Merry, 2016, p. 208; see also Kelley & Simmons, 2019). Such explanations offer critically important information on the current indicator frenzy and shed light on some of the underlying dynamics. Still, they account for only part of the story. In addition, they are liable to create path dependencies that ultimately impede our ability to gain a comprehensive understanding of GPIs: (a) The "magic of numbers" (Merry, 2016, p. 127) has become a conceptual shorthand and convenient catch-all formula. Extending the analytical scope, recent scholarship has argued that there is more to contemporary valuation devices than just numbers , which suggests that we should ask what other factors might account for the power that is frequently ascribed to GPIs. (b) By interpreting GPIs as (passive) carriers of global inequalities, scholarly debates gravitate toward the macro level and a "top-down" view. Some, though, have chosen a "bottom-up" strategy, asking how organizations facilitate global inequalities (Fehl & Freistein, 2020). In doing so, they grant the organizations that produce GPIs (GPI producers) a more active role (Ringel et al., 2020).

GPIs as Public Forms of Valuation
Studies on GPIs have documented changes in laws and regulations, the adoption of new standards, widespread reforms, and growing concerns about global challenges like public sector corruption, the quality of education, and climate change (Beaumont & Towns, 2021;Brankovic, 2021;Davis et al., 2012;Honig & Weaver, 2019;Hunter & Shaffer, 2022;Kelley, 2017;Kelley & Simmons, 2015Landahl, 2020;Wang & Rosenau, 2001). To further our understanding of these impacts, we must tease out "a broader focus on the wider sets of relationships through which [GPIs] emerge and remain authoritative" (Beaumont & Towns, 2021, p. 1469. It seems that this wider set of relationships comprises three main categories of actors: producers, targets, and publics (see also Samiolo & Mehrpouya, 2021;Waibel et al., 2021;Werron, 2014). Most producers of GPIs are organizations: international organizations, non-governmental organizations (NGOs), foundations, research centers, or, occasionally, for-profit companies. Nation states, the targets of GPIs, are evaluated in terms of their ability to perform in a chosen category such as health care, freedom of the press, and corruption control, or in any other tasks they are seen as fitted to fulfill. Lastly, GPIs also address publics, whom they present with the opportunity to evaluate the performances of nation-states. Unlike the producers and targets of GPIs, publics are a rather elusive concept, and defining them is a challenging task. In a very basic sense, publics emerges under the condition of what is generally called "publicity." In contrast to in-person conversations or private correspondence, public communication does not have a specific recipient but is, to use Brighenti's (2018, p. 25) term, "infrastructural"-once brought into existence, it takes on a life of its own and becomes "an 'air' that we breathe, an atmospheric component of society." There are different theories of "infrastructural" communication, each having direct implications for how we conceptualize publics. The Editorial of this thematic issue (Mende & Müller, in press) distinguishes four manifestations of publics: audiences, public spheres, institutions, and public interests. The second and fourth manifestations, public spheres and public interests, are key to unpacking the wider sets of relationships between producers, targets, and publics. Public spheres, and how they become arenas where publics interests are articulated, might be thought of in at least two different ways.
First, numerous studies approach public spheres as being "a concrete audience," even "a crowd witnessing itself in visible space" (Warner, 2002, p. 50). Accordingly, publics are smaller or larger groups of people who may be in proximity to one another or communicating at a distance. Modern nation-states have multiple publics: citizens, voters, the media, academics, lobbyists, civil society, corporations, etc. (Beaumont & Towns, 2021). GPIs assume the role of a "visibility agent" (Harness et al., 2022) that empowers these publics by transforming them into stakeholders who can evaluate nationstate performances, even if they only know little about the subject matter. For instance, most of us struggle when we try to grasp how modern health care systems work, but the Global Health Security Index offers an overall picture that radically reduces the asymmetry between health care experts and laypeople who are provided with the means to make up their own minds. The targets-government officials and civil servantsbecome aware of GPIs either when they follow public debates or because stakeholders confront them directly. Exposed to forms of valuation that cater to diverse publics, nation-states regularly engage in "status maintenance behavior" (Kelley, 2017, p. 52), which is also a popular theme in research.
There is another side to the emergence of today's vast indicator culture that has largely escaped scholarly attention: Contemporary GPIs not only put pressure on nation-states but also face challenges that ultimately endanger their producers' carefully built credibility as third parties (Hunter & Shaffer, 2022;Lokot & Wijermars, 2023;Nelken, 2015). On closer inspection, GPI critique, the flipside of GPI power, is a widespread phenomenon originating from different sources. Unsurprisingly, the representatives of nation-states consistently push back against GPIs (Merry, 2016). The media, too, act as promulgators of critique: journalists not only use GPIs in their reporting on nation-states but also comment on the measures themselves. The scientific community, another stakeholder, is just as much invested in calling GPIs methodologies and data quality into question as in supporting them. And finally, civil society also appears to be ambivalent: on the one hand, supporting and promoting GPIs while, on the other, finding faults in them (such as "complacency"). Like other nontraditional experts (Chong & Bourgoin, 2020;Ringel, 2021), GPI producers are evidently no strangers to contestation and put considerable effort into maintaining their evaluative credibility (Lokot & Wijermars, 2023). All things considered, the publicity of GPIs is a double-edged sword, putting producers and targets under pressure, thereby facilitating what I suggest calling the Janus face of valuation.
We can further explore this idea by turning to a second way of thinking about public spheres and public interests. This viewpoint is less interested in actors and instead foregrounds the larger discursive environment. Released in reports, newspaper articles, websites, launch events, etc., GPIs have an audience that, according to Werron (2014, p. 65), is "indefinite, unlimited, and thus basically unknowable." Not so much addressing specific groups of people but rather amorphous entities such as voters, patients, investors, or even mankind, GPIs enact "imaginations of the audience that project the audience as a 'public' of attentive and critical individuals rather than an unreceptive, undifferentiated 'mass'." It does not matter whether these imaginations have any correspondence to reality because a public sphere is "a space of discourse organized by nothing other than discourse itself" (Warner, 2002, p. 50). Instead of pondering what "real" preferences groups of people might hold, such conceptualizations of publics foreground their socially constructed nature. An increase in GPI power is therefore likely if the imaginations of publics as attentive observers of nation-state performance are spread in discourse. The same applies to GPI critique, which is likely to increase pressure on the producers to the extent that it can reach larger publics and successfully claim to speak on their behalf.
Valuation devices such as GPIs amplify what Power et al. (2009, p. 309) call reputational risk, which "can be regarded as the purest man-made risk of organizing as such, namely the risk of how one is perceived by others." Emerging from public discourse, reputation is an inherently social attribute awarded by others: no person or organization directly controls their own reputation. With the proliferation of highly diverse, and adverse, valuation practices in recent decades, we are witnessing a general trend toward "the internalization of concern about how activities might be regarded" (Power et al., 2009, p. 309). Organizations are, in other words, bound to expect reputational risk to be lurking around the corner so that GPIs likely wield more power over their targets and fuel more challenges to their producers' credibility by being able to create the impression that they could cause reputational risk.
Both understandings can help us get a better sense of the relational constellations in which GPIs are embedded. Conceptualizations of public spheres as sites of engagement between different actors reveal how stakeholders may use GPIs as a resource in their efforts to exert pressure-on targets or producers. Others conceptualize publics and their interests as socially constructed claims. Whether stakeholders "really" make use of them or not, GPIs generate pressure by virtue of their publicity. This means that GPIs, once embedded in public discourse, can be powerful irrespective of their utilization by stakeholders and, conversely, hold the potential to sway producers to act as if they were expecting challenges to chip away at their credibility. Treating both understandings as complementary, this article takes into account engagements between producers, targets, and publics as well as the more tacit pressure stemming from the placement of GPIs in public discourse.

GPI Power
GPIs have similarities with other public forms of valuation, especially rankings. Building on previous work (Ringel, 2021), this section defines GPIs as devices that derive their power from combining: (a) commensuration, (b) visual simplification, and (c) serialization, each element being built through the investment of considerable time, energy, and resources by the producers.

Commensuration: Making Nation-State Performance Comparable on a Common Metric
Commensuration is commonly defined as the social process of assembling entities-commodities, universities, corporations, people, and countless others-on a common metric (Espeland & Stevens, 2008). On the surface, commensuration may appear as unobtrusive and innocent, but research has shed light on its performative properties, suggesting that, rather than measuring something that is already out there, commensuration produces value, meaning that the relationship between estimation and value is "circular-better, entangled" (Brighenti, 2018, p. 25).
GPI producers devise different strategies to craft trustworthy numbers (Bandola-Gill et al., 2023). Over time, the producers who use primary data seem to have switched from previously informal and casual methods to a conspicuously formalized and standardized production process (Zerndt, 2020). Take the Access to Medicine Index by the Access to Medicine Foundation: According to a study, the team responsible for the index was constantly reminded that it must "act as a robot" (Mehrpouya & Samiolo, 2016, p. 22) to minimize (potential) accusations of human bias or error. GPIs that are either partly or fully built from secondary datacomposite indicators-aggregate different sources in "long interpretive chains" (Merry, 2016, p. 209) to tap into the credibility of these external sources. Whether primary or secondary sources are used, transparency has apparently evolved as a key source of legitimacy. Showing a marked interest in leveraging the cultural worth of "openness" by sharing (carefully prepared) information about their methodology and data, GPI producers resemble other nontraditional experts who utilize transparency to (pro)actively build trust because they do not possess the received authority of professional practice (Chong & Bourgoin, 2020).
The CPI is a composite indicator, commensurating (perceptions of) corruption on a global scale by aggregating several independent data sources. Being "very careful to appear neutral and noninterventionist" (Wang & Rosenau, 2001, p. 36), TI leans heavily on scientific principles when defining public sector corruption, deciding on methodological standards, or collecting and interpreting data, all of which is explained extensively in various documents to demonstrate the trustworthiness of the numbers in use. For example, the document "Frequently Asked Questions" (Transparency International, 2020) offers answers to questions such as: "What is the Corruption Perceptions Index (CPI)?"/"Which data sources are used for the CPI?"/"Why do we need the CPI if 13 other sources measure manifestations of corruption in the public sector?" TI also discusses its methodological decisions and how these decisions influence the scores. An update to the CPI methodology in 2012 is explained as follows: "From 2012, we will be using the raw scores from each of the data sources, which provide greater transparency as to how the CPI scores have been constructed" (Transparency International, 2012, pp. 1-2).

Visual Simplification: Presenting Valuations of Nation-State Performance in Appealing Formats
Numbers do not exist independently of the shape they are given. Espeland and Stevens (2008, p. 422) have argued that producers of numerical judgments show an inclination to take aesthetic matters into account as they craft "compelling, elegant, and even beauti-ful…numerical pictures." Studies on quantification that explore this theme have come to a similar conclusion, revealing that visual devices hold "communicative possibilities that are not found in the original [numerical] information" (Ronzani & Gatzweiler, 2022, p. 3; see also Pollock & D'Adderio, 2012).
Visualization is a key instrument in world politics (Freistein & Gadinger, 2022). In contrast to publications such as statistical yearbooks, which require high levels of numerical competency on the part of the reader, GPIs use aesthetic appeal to attract both expert and lay audiences (Bandola-Gill et al., 2021;Hansen & Flyverbom, 2015;Kelley & Simmons, 2019;Merry, 2016). Regardless of how sophisticated the terminology, (political) theories, methodologies, and data may be, a digestible presentational style makes GPIs entertaining and offers the audience an experience that is "more immediate" (Bandola-Gill et al., 2021, p. 34).
League tables erase ambiguities and nuances by visualizing an entire field of observation in clear-cut ranks (Davis et al., 2012). Across multiple sectors (higher education, tourism, professional sports, etc.), this format has emerged as the most common representation of hierarchical order . World maps are specifically used to visualize nation-state performances, usually by assigning colors based on a country's score, and play a major role in the public relations strategies of GPI producers (Bandola-Gill et al., 2021). Language, too, is used to achieve visual simplification. In a very basic sense, the label attached to a GPI already signifies "a simplification of what the index purports to measure or rank" (Davis et al., 2012, p. 75). Other examples are the punchy slogans and catchy "one-liners" which loom large on websites, in social media campaigns, PowerPoint slides, reports, and brochures (Seabrooke & Wigan, 2015).
The CPI is evidently geared toward visual simplification. The cover of the CPI 2021 report features an evocative picture of two hands exchanging money, which obviously alludes to bribery. In this CPI report, the index is introduced under the headline "180 countries. 180 scores. How does your country measure up?" (Transparency International, 2022a, p. 2), presented as a world map, and followed by an "executive summary" that links the scores to current world affairs, in this case, the pandemic: "Two years into the devastating Covid-19 pandemic, this year's Corruption Perceptions Index (CPI) reveals that corruption levels have stagnated worldwide" (Transparency International, 2022a, p. 4

Serialization: Transforming Nation-State Performance Into a Developing Property
The third element of GPI power is that the status conferred on "performers" is inevitably subject to change, resulting in careers in which they "climb," "stagnate," or "fall."  propose the term serialization to conceptualize the "close connection between time and social order" (Landahl, 2020, p. 627). Achieved through an elaborate balancing act that ensures stability while also preserving "the possibility of change" (Fourcade, 2016, p. 184), serialization, thus understood, depends on the ability of GPI producers to fulfill two distinct tasks. Only then can a "dry collection of numbers" (Landahl, 2020, p. 632) become a major event on the world stage (Brankovic, 2021).
The first task involves constructing narratives of performance as a volatile and elusive property that has to be measured repeatedly. For example, university rankings are rooted in the premise that institutional performance constantly increases or decreases; the same holds true for modern art rankings, which promotes the idea of a highly dynamic art market where "trends" abound . What distinguishes GPIs from these forms of quantitative valuation is an inherently global scope and a focus on a unique type of actor-the nation-state-that is thought of as continuously "devel-oping" (which is not to say "improving"). Heavily engaged in "scripting" (Auld & Morris, 2021, p. 186), GPI producers project imaginaries of volatile nation-state performances by crafting storylines that bring together diverse settings, events, and characters in reference to specific forms of temporality.
The second task involves putting the scripted storylines into practice by producing and publishing quantitative valuations on a regular basis so that each version is connected to past and (potential) future ones . Depending on the context, "regular" can mean very different things. In tourism, for example, the online platform TripAdvisor continuously updates scores, thereby accelerating the temporal order. Most GPIs establish annual publication cycles, though some are biennial, and a few (such as PISA) triennial. Because of the large number of GPIs being issued, finding a launch date can be very difficult. Moreover, GPIs face the challenge of addressing at one and the same time not only an integrated global system but also national political arenas and media debates (De Paola & Pirttilä-Backman, 2023).
Attending to both dimensions of serialization, the CPI generates high levels of attention (Andersson & Heywood, 2009;Wang & Rosenau, 2001). Corruption is framed as a developing property "out there," making periodic measurement and publication paramount. The 2021 report consistently invokes this frame. For instance, we are informed about the "most significant five-year movers" and told that "the CPI shows that control of corruption has stagnated or worsened in 86 percent of countries over the last decade" (Transparency International, 2022a, p. 7, emphasis added). By guaranteeing a year-by-year publication cycle and connecting present, past, and future scores in tables such as "CPI score changes, 2012-2021'' (Transparency International, 2022a, p. 7), TI enacts the CPI's basic storyline of corruption as being a global challenge that nation-states must tackle if they want to participate in the world's longterm improvement. TI also practices serialization on its website where an interactive map features a drop-down menu (allowing users to make temporal comparisons) and arrows next to country scores (indicating movement up or down).

GPI Critique
GPI power emerges from the combination of commensuration, visual simplification, and serialization. As we have seen, the producers are vital in this regard: they make commensuration possible (by devising "sound" methodologies and calculative practices); they craft visually appealing numerical pictures; and they implement serialized temporal orders. GPI critique is related to the same combination of elements that the producers also navigate by devoting considerable time, energy, and resources.

Commensuration: Fighting "Methodological Battles"
The extant literature has a penchant for connecting the proliferation of quantitative valuation devices to "the general social acceptance of practices of data collection and analysis" (Merry, 2016, p. 30). The same does not hold for GPIs: albeit numbers-based, their calculations almost never receive "general social acceptance." Different publics, most notably scholars, consistently give voice to methodological criticism and call the quality of data into question (Bandola-Gill et al., 2023;Nelken, 2015). Studies focusing on how the producers of quantitative valuations in other sectors typically respond to-methodology and data-related criticismhave revealed several strategies, among them the rejection of agency: university rankers, for example, position themselves as merely observing and measuring university performance, which, they claim, exists irrespective of what they do (Hamann & Ringel, 2023). GPI producers face greater difficulties in this regard. Their attempts to assume the role of disinterested monitors are apt to be viewed with suspicion on account of their embrace of a political agenda, such as the expansion of political freedoms, ending human trafficking, or improving education systems (Bandola-Gill et al., 2021;Merry, 2016;Seabrooke & Wigan, 2015). How, then, do GPI producers defend the "methodological veracity" (Beaumont & Towns, 2021, p. 1471) of their commensurated numbers when they are accused of lacking both the authority and neutrality that are necessary for this task?
One way in which GPI producers navigate (potential) "methodological battles" (Seabrooke & Wigan, 2015, p. 901) is by actively engaging in "consensus building" (Samiolo & Mehrpouya, 2021, p. 83; see also Bandola-Gill et al., 2023). They apply a range of inclusionary measures in the process of defining methodologies and gathering data to enroll "the audience in the means of qualifying the object" (Chong & Bourgoin, 2020, p. 89). To put it differently, publics are transformed into manageable stakeholders, the most important being the scientific community, civil society, and the public bureaucracy (elected officials, appointees, and civil servants): (a) Studies point to a growing number of interactions between GPI producers and scholars at conferences and workshops, where ample room is given for critical discussions revolving around methodological and data-related issues. These interactions become more permanent if scholars accept invitations from the producers to serve on boards or expert committees (Davis et al., 2012;Merry, 2016;Seabrooke & Wigan, 2015); (b) By engaging with NGOs, activists, and representatives of social movements, GPI producers ensure that politically speaking, they are on a solid ideological footing and cannot so easily be accused of having a "democratic deficit" (Beaumont & Towns, 2021, p. 1472); (c) GPI producers also foster relationships with elected officials, appointees, and public servants. Presented with an opportunity to make their voices heard by reviewing the data before their release, these stakeholders are more likely to accept being the object of evaluation-or at least refrain from open hostility (Honig & Weaver, 2019).
These and other types of seemingly neutral engagements effectively lure stakeholders onto a playing field that has been shaped by the GPI producers and is therefore tilted in their favor (Samiolo & Mehrpouya, 2021). Irrespective of any critical intentions they may have had in mind, once scholars, experts, NGOs, activists, elected officials, appointees, and public servants enter this playing field, they are bound to be co-opted as witnesses to the (promise of the) validity and soundness of GPIs. The CPI has been consistently criticized for its methodology and data (Andersson & Heywood, 2009). The following quote from an article published in Foreign Policy is just one of many examples: The problem with the index…can be found in the name. Perceptions are not facts, and in this case they may be an unhelpfully distorted reflection of the truth….The point here is not that any of these underlying sources are bad, or wrong, or anything other than what they claim. The point is that in aggregating them, the result lacks any sense of diversity. (Cobham, 2013) Exposed to challenges of this kind, TI relies heavily on its engagement with experts whose approval the organization constantly seeks and invokes in public statements (Seabrooke & Wigan, 2015). Until 2009 the economist Johann Graf Lambsdorff spoke for the CPI and, acting as the index's public face, provided extensive explanations as well as responses to critics, where necessary, in his role as a scholar. Another strategy applied by TI to defend the CPI is its dissociation from the data used for the production of the index, which, according to this line of argument, is collected by otherreputable-institutions: "The 2020 CPI draws on 13 data sources from 12 independent institutions specialising in governance and business climate analysis. The sources of information used for the CPI 2020 are based on data published in the previous two years" (Transparency International, 2020, p. 1).

Visual Simplification: Professing Nuance
Although visual simplification is an indispensable element in the success of contemporary GPIs, we should not neglect its propensity to trigger challenges. When critics refer to a lack of nuance and sobriety, what they usually have in mind are the streamlined messages, spectacular launch events, and bright colors in which GPIs are advertised. This reveals that visual simplification is only one side of the equation that must somehow be combined with testimonials to one's scientific credibility. But how is it possible to present "contested images of the world in simplified form" while at the same time professing "expert knowledge" (Broome et al., 2018, p. 529)? The literature points to different strategies by which these conflicting demands are navigated.
First, there seems to be a trend toward formalized, standardized, and progressively longer reports, issued both in print and digitally. These reports address informed and lay audiences alike (Zapp, 2022). Informed audiences of scholars, experts, policymakers, and administrators receive sophisticated methodological explanations and interpretations of the data, which underscore an organization's compliance with professional standards. Lay audiences (activists, journalists, interested citizens, etc.) are the main intended recipients of the executive summaries, bullet points, and graphical elements also to be found in reports as well as various videos, press statements, and social media campaigns. Owing to a certain interpretative flexibility and fuzziness, reports resemble boundary objects-artifacts that connect multiple social worlds and frames of reference. Given this multiplicity of purposes, it comes as no surprise that the preparation of reports is demanding and time-consuming (Zapp, 2022).
A second strategy for dealing with the conflicting demands of "visual simplification" and "professing nuance" is to downplay the narrative of winners and losers, whereby GPI producers "balance the clarity of message of the ranking and its political acceptability" (Bandola-Gill et al., 2021, p. 29). The world map is a prime example: by emphasizing areas or regions instead of individual performers, it avoids a "clear judgment on 'underperforming' countries" (Bandola-Gill et al., 2021, p. 41). In a similar vein, interactive maps entice the visitors of websites into becoming involved by allowing them to create individualized numerical pictures. This too alleviates the impression of a lack of scholarly sobriety and nuance.
Third, GPI producers strike a different note when engaging directly with some of their stakeholders after the publication. A study has shown that representatives of the previously mentioned Access to Medicine Foundation neglect the rank order of league tables in their engagement with the targets. Instead, they seek to enter into in-depth conversations about how the data could (and should) be interpreted in ways that reflect the full complexity of the findings (Samiolo & Mehrpouya, 2021).
The following quote from an article published in the Washington Post illustrates how the CPI's visual simplification is often subject to criticism: "It can be fun looking at tables like this" the author (Hough, 2016) admits, "and many people will indeed do precisely that over the next few days"--yet, readers are told, "we should be careful before reading too much into the data. The CPI is a particularly poor tool with which to judge a coun-try's anti-corruption successes and failures. Academic research highlights four reasons why." Preemptively countering charges to the effect that the CPI is nothing more than "fun," TI has turned into a prolific publisher of documents that are clearly designed to attest to the index's academic credentials. In its "Frequently Asked Questions" (Transparency International, 2020, p. 4), TI reminds us that the CPI, contrary to common misconceptions, "is limited in scope, capturing perceptions of the extent of corruption in the public sector, from the perspective of business people and country experts." In order to cover the "different aspects of corruption," the document continues, "Transparency International produces a range of both qualitative and quantitative research on corruption, both at the global level and at the national level.''

Serialization: Balancing Continuity and Discontinuity
Received wisdom in the social sciences has it that established GPIs face fewer challenges than new ones, which are more likely to receive criticism, whether directed at labels, methodologies, or data (Merry, 2016). The recent scandal surrounding the (as of now) terminated Ease of Doing Business Index, an example used by Davis et al. (2012) to illustrate the stability of serialized GPIs, suggests that this is not always the case. Moreover, a closer look reveals that any form of serialization requires constant care and attention, whatever the degree of institutionalization (Bowers & Prato, 2019). We should note that, compared to commensuration and visual simplification, critiques of serialization tend to be less frequent. But because GPI producers have to expect that serialization poses reputational risks, at least hypothetically, potential statements are of just as much consequence to them as actual ones.
As has been established, serialized GPIs stabilize a temporal order that makes nation-state performance a continually developing property: published at regular intervals, iterations are presented as being sufficiently similar to one another to appear intertwined. A periodically published GPI must argue that the 2011 edition follows directly from the edition of 2010 and leads on equally directly to that of 2012. Revisions, which usually affect the methodology, disrupt this sequential ordering. If, for example, the 2011 edition uses different data or devises new calculative practices, it ceases to be the direct successor of 2010 and, as a result, serialization is suspended. Why would GPI producers choose to imperil the temporal rhythm that they have so carefully crafted?
There is reason to believe, for one thing, that some methodological revisions are made in the pursuit of newsworthiness. As a study on corporate rankings has suggested, it is not enough to ensure regular publication, scores should also show sufficient movement in the projected ordinal scale to sustain an audience's interest (Bowers & Prato, 2019). This seems to apply to GPIs as well (Beaumont & Towns, 2021;Brankovic, 2021): scores that remain similar over the years can easily be perceived as "boring" because they are "too predictable." A study of PISA, for example, emphasizes the necessity of presenting "an unpredictable and changing hierarchy of educational systems…that conveys the image of a world in flux…of decline and development, of rising and falling stars" (Landahl, 2020, p. 633).
Validity is another reason for revisions that disrupt serialization. Sometimes, circumstantial changes such as legal reforms or the availability of new data make updates necessary (Merry, 2016). In other cases, revelations of methodological flaws put pressure on GPI producers who subsequently try to exploit revisions by invoking them in their public communication, where they claim that past and present flaws have been or will be rectified (Ringel, 2021). Under these conditions, rather than posing a threat, "continuous scientific debate and refinement remains part of the indicator's further life" (Davis et al., 2012, p. 89). What is more, critics are enlisted into a collective project of ultimate improvement to which their "constructive criticism" might contribute, whether they intended it to or not.
Serialized GPIs depend on their producers' skills in navigating continuity and discontinuity. But thus far, research has shown only minor interest in how these conflicting demands are tackled. One option is to recalculate the scores of previous iterations, which, depending on the methodology and staff available, can be quite costly and complicated. Alternatively, GPI producers could practice "organized hypocrisy" (Brunsson, 1989) by decoupling what they say ("talk") from what they do ("action") in order to survive in increasingly complex environments (Ringel, 2021). While perhaps more efficient in terms of costs and time, this strategy is delicate in a different way: it makes the producers vulnerable to charges of inconsistencies between "talk" and "action." TI has attended to the task of balancing continuity and discontinuity throughout the history of the CPI. As described in the previous section, different iterations are skillfully interwoven in a seamless web of measurements that transform corruption into a continually developing property. At the same time, the methodology has been subject to change. Until 2009 revisions were planned and executed by Johann Graf Lambsdorff. But, as TI was forced to realize, relying on one person comes with a price attached. In 2009, Lambsdorff denounced the CPI in an e-mail published by Global Integrity (Global Integrity, 2009) and called for the index's termination: "TI-S [TI Secretariat] will try to continue somehow with a substitute for the CPI," he suspected, which is why he thought it "time for me to let them go their way. From now on, it is only TI-S which will sign responsible for the outcome. I won't be out there to take the honor, nor the blame" (Global Integrity, 2009). He further made it clear that "I won't be out their [sic] to provide academic credibility and link the data to an international research agenda that hast [sic] swept through all major scientific journals" (Global Integrity, 2009).
Lo and behold, the CPI survived the attack-by leveraging the opportunities of methodological reform. A document published in 2012 explains that the latest edition "has been calculated using an updated methodology," which TI "developed following a comprehensive review and consultation process, both within the Transparency International movement and with the help of external experts. The updated methodology has been reviewed and validated by an independent statistical assessment" (Transparency International, 2012, p. 1). The document also highlights that "due to the update in the methodology, 2011 CPI scores are not comparable with CPI 2012 scores….Scores from the CPI 2011 and previous editions should not be compared with scores from 2012." TI has reaffirmed this stipulation over the years, for instance in its "Frequently Asked Questions," where readers are reminded that "due to a change in our methodology in 2012, results from before that year cannot be compared. Only CPI results from 2012 onwards can be compared" (Transparency International, 2020, p. 3).
Among the strategies at its disposal to navigate continuity and discontinuity, TI apparently favors "organized hypocrisy." Statements such as the above, which seek to discourage comparisons between the editions published before and after 2012, affirm the scholarly standards of (critical) expert audiences. This, however, stands in contrast to the interactive map on the TI website, which is obviously designed to be consumed by (larger) lay audiences. Devoid of the nuances so vividly put on display in methodological statements, the interactive map's drop-down menu mentioned earlier presents an uninterrupted temporal sequence from 1995 (the publication of the first CPI) until the latest edition, including, of course, scores from before and after 2012.

Conclusion
There is ample evidence that GPIs have a profound impact on the world. They prompt changes in laws and regulations, spread new standards, create favorable conditions for widespread reforms, and bring awareness to global challenges. What are the main drivers of today's vast "indicator culture" (Merry, 2016, p. 9)? Showing a preference for structural explanations, social science research tends to contextualize the proliferation of GPIs in larger transformations, notably the institutionalization of trust in numbers and global inequalities.
Synthesizing the pieces of literature on quantification, publics, and valuation as well as empirical studies on GPIs, this article has shed light on other factors. GPIs were conceptualized as Janus-faced valuation devices because the very sources of their power-commensuration, visual simplification, and serialization-also fuel challenges to their credibility. The organizations that produce GPIs dedicate a great deal of time, energy, and resources to assuming the role of "visibility agents" (Harness et al., 2022) who address, and speak on behalf of, publics. In this capacity, they fulfill the intertwined tasks of building GPI power and navigating GPI critique. A major implication of this conceptual argument is that structural explanations such as a trust in numbers and global inequalities, although providing vital insights, only tell half the story. As for the trust in numbers, we have seen that visual simplification and serialization are just as important as the calculative practices themselves. And, despite perhaps profiting from the favorable "top-down" conditions of global inequalities, the producers are deeply invested in sustaining GPIs "bottom-up." The general thrust of this article was towards a general perspective. It is therefore essential that future studies provide in-depth knowledge on variations between GPIs and how their producers deal with the tasks of building GPI power and navigating GPI critique. Variations may be studied concerning the three elements: commensuration, visual simplification, and serialization.
Each type of commensuration comes with its own affordances and constraints. The definition of "nationstate performance," for example, is a dimension in which we can expect intriguing variations, both substantively and analytically. Substantively, there is a staggering variety of nation-state performances that GPIs claim to measure, which begs the question of why some gain traction and become powerful, despite still being criticized, while others are simply ignored or meet such a degree of resistance that they ultimately fail (Lokot & Wijermars, 2023). Analytically speaking, most GPIs emphasize positive characteristics such as who has the best healthcare system, who can deliver quality education, and so forth. Yet there are a select few cases of GPIs that measure and highlight negative performances, with the Global Slavery Index being a prime example (Brankovic, 2021). The logic of ordering is another dimension of interest (Towns & Rumelili, 2017). Some GPIs, such as the Human Development Index, operate according to a zero-sum logic: ranks of nation-states come at the expense of their competitors. Others, such as the Trafficking in Persons Report, are made into ratings or classifications, which amount to hierarchies between classes or categories rather than individual performers. Finally, special attention should be given to GPIs that have switched from one type of order to another (Hunter & Shaffer, 2022).
Because GPI producers make extensive use of visual simplification when they try to reach larger publics, they increase the risk of being criticized for a lack of sobriety and nuance. Previous research has provided some indications that global reports-a very common type of publication-are a valuable empirical source that reveals directly how GPI power and GPI critique are navigated in practice (Ringel, 2021;Zapp, 2022). Building on these findings, future studies could provide in-depth ethnographic accounts of how reports are crafted, thereby opening the black box of GPI production (Mennicken & Espeland, 2019). Alternatively, reports could be analyzed as manifestations of discourse and sites where narratives are promoted (Auld & Morris, 2021).
By virtue of serialization, GPIs create a fluid status order that puts the target's reputation at risk. The producers deal with accusations related to their impact in different ways. Notably, some GPIs are explicitly designed to intervene while others, such as the CPI, take great care to remove themselves from the picture by assuming the role of the disinterested observer. The issue at hand is then not if GPIs have an impact but whether or not their producers admit to having agency and embrace their role by verbalizing and promoting specific notions of public interests (Mehrpouya & Samiolo, 2016). Among those who explicitly see their role as being facilitators of public interests, we can distinguish between "reformist" and "revolutionary" GPIs (Seabrooke & Wigan, 2015). Reformist GPIs are moderate; they seek incremental change within agreed-upon parameters. Revolutionary GPIs, by contrast, seek radical change whereby they fundamentally question established orders of worth. We might expect reformist and revolutionary GPIs to use different strategies when tackling the tasks of building serialized GPI power and navigating serialization-related GPI critique.