Facts, power and global evidence: a new empire of truth

Abstract What are the epistemological and political contours of evidence today? This introduction to the special issue lays out key shifts in the contemporary politics of knowledge and describes the collective contribution of the six papers as an articulation of what we describe as a ‘new empiricism’, exploring how earlier historical appeals to evidence to defend political power and decision-making both chime with and differ from those of the contemporary era. We outline some emerging empirical frontiers in the study of instruments of calculation, from the evolution of the randomized controlled trial (RCT) to the growing importance of big data, and explore how these methodological transformations intersect with the alleged crisis of expertise in the ‘post-truth’ era. In so doing, we suggest that the ambiguity of evidence can be a powerful tool in itself, and we relate this ambiguity to the ideological commitment and moral fervour that is elicited through appeals to, and the performance of, evaluation.


Introduction
Instead of considering reason as the truth of the subject, we can look to other foundational categories that are less abstract and more tactile, such as life and death. (Achille Mbembe, Necropolitics, 2003, p. 14) Today the authority of facts is both far-reaching and fragile. The growing influence of randomized controlled trials (RCTs) in economics, political science, development studies and humanitarian work underscores the dominance of an 'evidence-based' style of decision-making in contemporary governance. Global health, a field that has quickly consolidated around econometric models and epidemiological indicators, has effectively transformed health care into a data management enterprise. Think-tanks and private 'economic intelligence' organizations proffering independent research to guide social and economic policymaking abound. On the cusp of the 'Big Data Revolution', largescale calculations of population behaviour are seen as an elixir promising to revive the economy, modernize public service delivery and enhance the impact of emergency aid. And yet, in the face of this muscular form of factmaking, faith in the technocrat is flagging.
Those who would believe that we have entered into an age of 'post-truth' politics are certainly not short on examples. The erroneous economic calculations and wild diversity of quantitative guestimates that have dominated the debate over Brexit in the United Kingdom; the 'fake news', 'alternative facts' and outright lies that defined the rise to power of the current American president: these feel like watershed moments in the abandonment of cherished ideals about the role of factual evidence in the conduct of political debate. For those committed to a 'marketplace of ideas', the apparent disinterest, on the part of both political leaders and their publics, in parsing the veracity of political statements is particularly disheartening. Pundits blame economic disenfranchisement, anti-elite sentiment and a desire for change; expert opinions that support the status quo are commonly rejected, regardless of the accuracy of their claims. The triumph of Britain's Leave Campaign and Donald Trump's ascendancy lay in grasping the popular zeitgeist and providing it with a plausible scapegoat, whether in the figure of the illegal immigrant, the climate change ideologue or the faceless EU bureaucrat. Though roundly discredited by armies of historians, policy-analysts and fact-checkers, these 'truths' drew their rhetorical power not from their technical accuracy, but rather from their emotional salience (Povinelli, 2017). Unconstrained by the need to substantiate arguments with extensive research or reliable figures, political discourse, many fear, has become a race to the bottom.
What constitutes authoritative evidence in this political climate? To what uses is evidence put, and what values does it carry? What obligations must be placed on the companies, such as Google or Facebook, that configure our new public spheres while profiting from the tracking and steering of online behaviour?
What counts in the making of facts, and who does the counting? Which empirical tools and metrics garner sufficient political capital to guide policy during times of economic uncertainty? And, critically, how do the social sciences respond to the increasing social and political significance of data while accounting for the deepening popular scepticism of the facts that data are used to support?
These questions demand cross-disciplinary reflection. Recent scholarship engaged with the social and epistemological implications of new forms of data collection, including virtual worlds and new valuation metrics, offers a critical starting point to reconsider the contemporary political contours of empiricism (Adkins & Lury, 2011;Espeland & Sauder, 2007;Fourcade, 2011). New digital tools provide a focal point to these discussions, with some commentators energized by the potential of aggregated data sets to disrupt the distinctions between qualitative and quantitative approaches, generating richer insight into political movements, economic transformations and social configurations (Latour et al., 2012;Marres & Gerlitz, 2016;Savage & Burrow, 2014). Others are less sanguine, emphasizing the degree of corporate control over 'big data', as well as the capacity of technocrats to 'reduce the noise of context', deriving managerialist solutions from aggregations of abstracted information (Biehl, 2016, p. 129; see also Adams, 2016;Tufecki, 2014).
The papers in this special issue offer innovative ways to conceptualize what we will describe as a 'new empiricism' using examples from global health, international development, humanitarian philanthropy and climate change policyfields which derive their political support and public salience from appeals to emergency, social justice and environmental stewardship. Authored by Donovan, Ehrenstein, Kelly, McGoey, Neyland, Reubi and Thiel, the pieces draw together different critical traditions, but share an abiding concern with the scaling, pace and symbolic power of fact-making, and what the temporal and spatial dimensions of evidence-gathering mean for the shifting relationships between knowledge, ignorance and power today.
Of particular interest is the conceptual strain that the purported 'crisis of expertise' places on empiricism and, more broadly, on political theory. Below, we elaborate the histories of several evidence-based discourses, providing new insights into the distributional logics and welfare claims implicit within the RCT revolution. Our analysis focuses on the longue durée, rooting the 'new empiricism' in the upheavals that marked the end of the Enlightenmenta period when a cross-section of political theorists, including Edmund Burke and, later on, Alexis de Tocqueville, insisted upon the need to base governmental decisions on a firm command of available evidence, to develop a 'science of government', in Burke's words (1968Burke's words ( [1790. Notably, the faith in evidence expressed at the end of the eighteenth century was punctuated with theological claimsto defend royal prerogative or the right of financial inheritance, in Burke's case, or to justify imperial conquest, in Tocqueville's. In Burke's case, he drew on God to defend property rights as he denounces the French revolutionaries' claim to a 'right to food' (Burke, 1790, p. 52). Tocqueville appealed to God's authority to explain how, against all odds, the world's weakest and most impoverished peoples had unshackled themselves from feudal rule. In Democracy in America, he writes that the 'magnitude of present achievement' in achieving more equality suggests that the 'gradual progress of equality is something fated' by divine will (Tocqueville, 1966(Tocqueville, [1835, p. 12). Later on, he also references God's authority when defending his staunch support for France's colonial conquest of Algeria. Tocqueville's and Burke's range of rhetorical registers illustrates the ease with which leading Enlightenment thinkers were able to draw on competing and often contradictory logics to legitimize established forms of aristocratic privilege, colonial wealth extraction and white entitlement.
The ability to switch between 'theological' and 'empirical' justifications for democratic governance, on the one hand, and imperial conquest, racial subjugation and class domination, on the other hand, has not abated among elite figures today, even if social scientists have yet to fully explore the enduring pertinence and centrality of theology to twenty-first century capitalist expansion and social stratification (Blencowe, 2016;Comaroff, 2009;Cooper, 2011;Cooper & Graham, 2015;James, 2015;Konings, 2015;McGoey & Thiel, 2018). Concerned with the ways that evidence can act as both a form of 'ecstatic knowledge', in Kelly's phrasing, and a form of 'charismatic violence', as McGoey and Thiel suggest, our contributions call for a sociology of the 'new theology' to complement explorations of the 'new empiricism'.

Empires of post-truth
Are we living in a post-truth era? Negotiating the cleavages between truth and political opinion was, according to Hannah Arendt, both the cost and the duty of a pluralistic society. If, in accordance with democratic principles, we value the free expression of a diversity of views, then we must be willing to accommodate disagreement on the nature of human affairs and how best to organize them. Levied for political purposes, facts convey a sense of self-evidence that preclude debate and pre-empt civic consensus. 'Truth', Arendt writes, 'has a despotic character ' (1967, p. 54). Yet, at the same time, it is factual truthor rather the notion that facts exist independent of our beliefsthat gives democracy its purpose and orientation. Respect for 'the indestructibility … of brutally elementary data' is what differentiates the demagogue from the spin-doctor and democratic states from totalitarian ones (Arendt, 1967, p. 53).
Today the sources of data have multiplied and fragmented, reviving old questions and raising new ones about the trustworthiness of empirical facts, their provenance, their claim to objectivity and their ability to reveal underlying patterns of causation (Levin & Leonelli, 2017;Parry & Greenhough, 2017;Prainsack & Buyx, 2017). The explosion of digital media and smart technologies have radically transformed the ease, scale and speed with which data are generated and circulateda quantitative deluge often unbound by expert judgement or experimental design. The ever-widening gap between numerical data and a meaningful fact is something that critical theorists see as the hallmark of our contemporary 'data society'a world in which adjudicating the legitimacy of political opinion is secondary to gauging and shaping public sentiment for commercial gain (Davies, 2016a;Ehrenstein, 2018;Leonelli, 2015;Lezaun, 2013). When the algorithms and analytical instruments that inform major news outlets, political pollsters and multinational companies are those trafficked by 'computational propagandists', it hardly seems surprising that scepticism about the authority of experts is on the rise (see Woolley & Howard, 2017).
The roots of post-truth, however, pre-date the digital revolution. Since at least the late Enlightenment period, the inherent contradictions between a valorization of self-interest and private property, on the one hand, and a commitment to equality and public good, on the other, constitute the fault-lines around which democratic politics are organized. At the end of the eighteenth century, Burke appealed to the righteousness of empirical evidence and the superiority of rule by experts to contest the notion of a democratic right to food, liberty or human equality. 'What is the use of discussing a man's abstract right to food or medicine?' Burke remarks in his 1790 essay Reflections on the revolution in France, 'The question concerns how to procure and administer them. In that deliberation I shall always advise calling in the aid of the farmer and the physician rather than the professor of metaphysics' (Burke, 1968(Burke, [1790, p. 151).
Burke's defence of aristocratic privilege and rule by experts was contested by pro-democracy adherents such as Mary Wollstonecraft and Thomas Paine, who each issued manifestos in defence of the 'rights of man' within months of the publication of Burke's pamphlet. They refuted Burke's insistence that aristocratic rulers could judiciously allocate resources when the recipients were deemed to be of subhuman worth and value in comparison to the rulers. Competing presumptions of natural equality, in Wollstonecraft's and Paine' views, and the natural inequality of men, in Burke's view, underpinned conflicting views on the legitimacy of French revolutionaries' right to receive a fair share of their nation's wealth. A little over two centuries later, similar debates over the right to economic and social security underpins populist frustration over growing wealth inequality, leading to competing truths over the propriety and defensibility of current global wealth disparities.
Today, the paradoxes of aligning private commercial interests with public welfare notwithstanding, the notion that maximizing private profits is somehow fundamental to human nature is precisely the kind of despotic truth that Arendt warns against. Neoliberalism, a radical extension of the liberal tradition associated with scholars such as Tocqueville, operates through a sort of Arendtian 'brutally elementary fact': the assumption that competition is the most efficient and thus incontestably advantageous way to allocate public goods (e.g. Crouch, 2011; Davies, 2016b;Harcourt, 2011;Ong, 2006;Read, 2009).
Under neoliberalism, therefore, rational policy does not arise from a deliberative consensus over how best to serve public interest, but rather from the seemingly objective need to enhance market efficiency to promote economic development (e.g. Boltanski & Chiapello, 2005;Cooper, 2011;Lazzarato, 2009). The cascade of policies undertaken in the name of value for moneyfrom deregulation and tariff elimination to the privatization of public domains previously considered exempt from market relationshave dismantled social safety nets intended to protect the most vulnerable in society and radically exacerbating social stratification and inequalities (Mirowski, 2013;Peck, 2008). The universalizing logic of cost calculation evacuates the politics from these policies and discredits alternatives as failing to meet the principles of good management (Clarke, 2004;Power, 1997).
Thus another take on the rise of populismone considerably more compelling than the expression of mass confusion perpetuated by internet bots and blowhardsis that it represents a rejection of the neoliberal thesis, and with it all expressions of its 'rational' logic. The spectre of corporate capture haunts purveyors of evidence-based policies and good governance on either side of the left-right divide. Indeed, if anything is to be learned from the 2007-8 financial crisis, it was that economists, their complex mathematic models and speculative projections misrepresent and ultimately fail to understand the realities of the economy (MacKenzie, 2009;Muniesa, 2014). The social and political failures of neoliberalism, in short, have profound epistemic reverberations: post-truth is managerialism come home to roost (Clarke & Newman, 1997;King, 1999).
The question then is: where does a sociological examination of the authority of evidence go from here? Materialist critiques of neoliberal ideology have been productive in illuminating the forms of domination and oppression obscured by economic discourse (Harvey, 2007;Sunder Rajan, 2006). Foucauldian analyses have served to further particularize and historically situate those dynamics of power and knowledge and to provide insight into the aesthetics of neoliberal governance and the varied subjectivities to which contemporary 'regimes of truth' give rise (Barry et al., 1996;Collier, 2012;Miller & Rose, 2008;Rose, 1999).
But there is a palpable sense too of academic frustration with the limits of critique, perhaps illustrating a growing adoption of what Hirschman (1991) called the 'perversity' thesis: in this case, the idea that progressive critiques of the status quo unwittingly reinforce the power of conservative forces by calling into question the epistemic foundations that have generally supported policies for positive social change. Living in the era of 'post-truth politics', critical scrutiny of the politics of truth can thus feel at once overdetermined and theoretically underpowered. The role of post-structuralist theory in 'the dissolution of the modern fact' has prompted much self-reflection from scholars of science and technology (Collins et al., 2017;Lynch, 2017;Sismondo, 2017, p. 4). While some read the popular rejection of expertise as evidence of a much-needed epistemic democratization (Fuller, 2017), others caution against the authoritarian implications of a wholesale destruction of scientific norms and institutions (Baker & Oreskes, 2017;Phillips, 2017).
Anxieties of responsibility and fears of conceptual capture pervade these discussionsa shadow cast more clearly over the climate change debate, where description of the contingency, work and, critically, politics involved in the making of evidentiary claims can be easily used to bolster scientific denialism (Demeritt, 2006;Edwards, 2010; see also Fitzgerald, 2017). But beyond taking sides in a burgeoning 'war' between scientific integrity (Latour, 2017) and the abandonment of truth to expediency, might there be other critical avenues to follow?
From post-truth to more truth One way forward may be to recognize the fact that growing political challenges to the status quo do not necessarily herald a reactionary, anti-enlightened flight from truth, but rather a demand for more truth. The deeper truths that surface at the fraying of science and authority occasion a radical form of critique (see Caduff, 2014)an unblinking assessment of the racial politics, colonial and neocolonial violence that provide the warp and weave of liberal democracy. We have entered into a 'negative moment', as Cameroonian historian and philosopher Achille Mbembe puts it, a period 'when contradictory forcesinchoate, fractured, fragmentedare at work but what might come of their interaction is anything but certain' (Mbembe, 2015).
The forces Mbembe charts emerge from the field of modern state intervention conceptualized by Foucault as biopower. However, rather than elaborate the rationalities, techniques and practices intended to maximize the life of the population, Mbembe's analysis begins from the extension of governmental control over death. Politics, he argues, is an art of differentiation: dictating who may live and who may die, which bodies have value and which are expendable. The 'romance' of personal autonomy that is the mainstay of liberal politics obscures the 'instrumentalisation of human existence and the material destruction of human bodies and populations' that grounds the sovereign project (Mbembe, 2003, p. 14). Reading back into Homo economicus the logics of the slave plantation, Mbembe illuminates 'the relations of enmity'the hostilities, immiseration, expressions of hate and organized killing necessary for the democratic project to be realized (Mbembe, 2016).
One of the heuristic advantages of a 'negative' reading of liberal democracy is that it allows us to better parse the analytical importance and, critically, the 'promiscuous pervasiveness' of the rise of so-called populism in political practice and discourse. 1 As Molyneux and Osborne (2017) suggest, the term populism has become a convenient placeholder for the varied and competing demands of groups with discordant grievances, some of whose complaints are more legitimate than others. 'More often than not when the concept of populism is invoked', they suggest, 'we might be invoking with greater pertinence other phenomena such as authoritarianism, Caesarism, illiberalism, class struggle, political romanticism, racism and so on' (Molyneux & Osborne, 2017, p. 17). This is an important point, but their short paper does not develop a full study of the cost of these occlusions, nor highlight earlier efforts from critical race theorists to insist upon a closer study of liberalism's Janus face.
The sociologist Paul Gilroy points out similar methodological occlusions in Foucault's work: the tendency to track the history of particular knowledges in a way that inadvertently devalues and thus obstructs attention to the construction of 'truths' he leaves unperturbed. One example is Foucault's comparative inattention to the racist logic of colonial conquest. In Gilroy's words: 'The extensive debate as to whether Negroes should be accorded membership of the family of mankind … was more central to the formulation and the reproduction of the modern episteme than Foucault appreciated. I raise this, not to pillory him … but because his study of that fateful change is an important resource in our own situation where similar processes are observable' (1998, p. 847). 2 From the purview of the slave trade and imperial subjugation of nations deemed by J.S. Mill and his fellow colonial thinkers as too 'semi-barbarous' for self-governance (Mill, 2015, p. 405), suspensions of law and juridical order are not modernity's state of exception but its rule. A genealogy of liberalism that begins from colonialism spotlights the forms of apartheid that structure the 'free market': catastrophe bonds, vulture funds, microfinancing and other systems that transform debt, destruction and crisis into forms of investment are particularly vital arenas for a necropolitical analysis. Rejoining liberalism to empire resets racism as the orienting ideology of modern governmental reason (Bhambra, 2016(Bhambra, , 2017Kish & Leroy, 2015;Mehta, 1999;Seamster & Charron-Chénier, 2017). In short, administrative strategies, bureaucratic calculations and judicial institutionswhat Judith Butler terms the 'epistemological ruse of power'cannot be disentangled from the efforts to denigrate the lives of one population in order to secure the dominance of another (Butler, 1992, p. 67).
One of the central lessons of histories of European colonialism was that the enterprise was beset by farce and futility, ignorance and inaction (e.g. Fabian, 2000;Hunt, 1990;Vaughan, 1991). However, rather than reading the limitations of colonial rule in terms of an administrative failure, one must situate the art of government within a logic that encompasses its political and emotional failures and excesses (Lachenal, 2017;Stoler, 2010). Driven by 'anxieties of annihilation' and 'fantasies of extermination', colonial logic, Mbembe argues, was 'simultaneously religious, mystical, messianic and utopian' (Mbembe, 2011, p. 116). The will and sentiments of rule, the collective desires and destructive exuberance, are the pillars of modernity, the categories that, Mbembe suggests, form the basis of a genealogy of truth.
This special issue takes up Mbembe's call to develop categories for a critique of reason that are 'more tactile' in two ways. First, it seeks to analyse the epistemological assumptions underpinning the growing emphasis on evidence-based discourses in contemporary policymaking, training our attention on the normative, affective and aesthetic dimensions of evidence's epistemic authority. The legitimacy of the facts discussed across the special issue hinges upon the vividness with which they conjure suffering and evoke salvation. Giving that rhetorical force empirical traction, this collection helps to excavate the 'structures of feeling' that subtend contemporary evidentiary claims and provide new conceptual tools for sociological critique (Williams, 1977).
Second, in pursuing how the production of sentiment in the context of these evidence-based initiatives structures the 'distribution of the sensible' (Rancière, 2004), the six papers offer new ways of conceptualizing the political economy of truth. Bringing together anthropologists, sociologists and geographers, the special issue explores why certain forms of evidence are valued over others; which populations and voices are strengthened or silenced in the making of new forms of evidence; and how political paradigms embrace or disavow the value of evidence in shaping new understandings of causal relationships, social obligations and political positions. We thus hope to take this 'negative moment' as an occasion to confront the feeling of factsto elaborate the epistemic and aesthetic grammars that cut across the bureaucrat and the expert, the troll and the global activist.

The moral authority of evidence
The faith of the current era in the power of measurement and quantification as an instrument of government is, of course, hardly new. From Tocqueville's writing on the effort to build a 'new science of politics' (1835, p. 12) in nineteenth-century America, to utilitarian efforts to promote the common good through the development of algorithms for calculating the costs of public and private risks, the modern era has often been defined by what the political theorist Letwin (1965) described as the 'pursuit of certainty': the effort to restrict ideological influences on government decisions by basing policies on evidence. And yet, as Porter (1996) suggests, a society's penchant for placing trust in numbers and other tools of standardization is rarely itself an inevitable or 'standard' phenomenon: cultures often differ dramatically in the types of evidence they commission or censor, revere or denounce.
The empirical imprimatur of modern democracy is the liberal orientation of knowledge: the degree to which knowledge can be freed from authoritarian judgement. Indeed, when Ronald Aylmer Fisher developed the method of randomized design that would form the basis of the RCT, he saw its revolutionary potential in generating results that 'any thinking man could understand' (Fisher, 1935, p. 65). However, drawing conclusions from randomized experiments was never as straightforward as bearing witness to their outcomes; the early days of the RCT were marked by strident debates on the terms through which statistical significance ought to be adjudicated (Kelly, 2011). Contemporaries of Fisher like Jerzy Neyman and Egon Pearson, for instance, advanced a decision-oriented statistical framework that foregrounded the subjective nature of judgement by emphasizing that the degree of confidence required by an experimenter depended on the ends to which an outcome was put. Neyman and Pearson regarded inference not as a matter of knowing the truth, but rather 'an act of will … to take a particular action, perhaps to assume a particular attitude towards the various sets of hypotheses' (Gigerenzer et al., 1989, p. 99). Fisher, for his part, claimed that their model entailed consequences 'horrifying for intellectual freedom in the west' (Fisher, 1955, p. 70).
A rhetorical intensity at odds with the apparent neutrality of statistical techniques, these debates underscore just what is at stake in the rapprochement of authoritative knowledge and decision-making in real-world experiments. RCTs promise to provide a clear answer to the question of 'what works', 'unsullied', as Donovan (2018) puts it, 'by ideology, politics and fads ' (2018, p. 15). Their apparent evidentiary robustness serves as a moral anchor, weighting even the most uncertain of evidence claims with an aura of equitable evaluation and distributive justice. And yet, despite the appeal of evidence freed from subjective bias and contextual contingencies, the epistemic blind-spots of an RCT hegemony has been the subject of considerable controversy. Medical practitioners such as Iain Chalmers and Ben Goldacre condemn commercial secrecy laws that permit companies to bury negative trial results, thus endangering lives (see Chalmers, 1990;Jørgensen et al., 2006;Kendall et al., 2009), while others express concern about the appropriateness of trialled interventions when the problems that are being addressed reflect broader socio-economic structural inequalities that singular interventions are unlikely to removeto name just a few of the limits of RCT methodologies which Donovan's paper in this special issue further unpacks (cf. Deaton, 2009;Donovan, 2018;Kelly & Geissler, 2012;Petryna, 2007;Timmermans & Berg, 2003;Wahlberg & McGoey, 2007;Will & Moreira, 2016).

From gold standard to old standard
We suggest that the evidentiary authority of RCTs to enable governing authorities to make fair and prudent decisions about research allocation is both the triumph and the failure of this methodology. In the triumphalist corner, there are the 'randomistas', a powerful epistemic community who share normative faith in the value of randomized 'field experiments' to provide causal evidence of a policy's effectiveness. Massachusetts Institute of Technology economists Abhijit Banerjee and Esther Duflo have been at the forefront of these efforts. In their bestselling book, Poor economics, they cite George Orwell to bolster a more empathetic approach to the study of poverty and economic inequality: it is only 'natural' for the poor to spend money on seeming 'indulgences', because it allows them to rationally survive hardship without losing their sense of dignity (Banerjee & Duflo, 2011, p. 38).
Their reference to Orwell has an earlier precedent: Adam Smith's famous 'linen shirt' example, thought to be one of the earliest modern illustrations of the importance of using empirical evidence to reduce both absolute and relative poverty. Smith's point is a compelling one. In Wealth of nations, he acknowledges that 'strictly speaking' a linen shift is 'not a necessary of life', before adding that in 'the present times, through the greater part of Europe, a creditable day-labourer would be ashamed to appear in public without a linen shirt, the want of which would be supposed to denote that disgraceful degree of poverty which, it is presumed, nobody can well fall into without extreme bad conduct' (Smith, 1999(Smith, [1776, p. 465).
To lack a linen shirt was thus seen as furnishing in the eyes of both the labourer and his superior clear and irrefutable evidence of the former's moral failings. It is a type of 'tactile fact' that no working man could afford to ignore, for without the means to appear respectable no worker could prove that he was deserving of receiving more work. Banerjee and Duflo seize upon what we see as the primary meaning conveyed by scholars like Orwell and Smith: that even if, say, a shirt made of fine cloth does not at first seem as essential to survival as, say, food, water or shelter, appropriate clothing is critical to avoid shame and the stigma of being impoverished. Smith's linen shirt or Orwell's 'cheap luxuries' are no different than a television or a cell phone that people in poverty are sometimes maligned for desiring. Wants, Duflo and Banerjee insist, can be just as important for economic survival as 'needs'.
This is an important point, and it is widely shared among the new randomistas. But we suggest their approach often neglects the secondary meaning of Smith's example. Smith is not simply making an argument about relative poverty. Rather, he is making an ever more profound assertion about external judgement and the problem of social bias. The lack of a clean shirt, he points out, is seen as evidence of moral failure because 'it is presumed' that 'nobody can well fall into [poverty] without extreme bad conduct'. Note that Smith does not agree that individual misconduct or immorality alone can lead to personal penury. 3 He is simply conceding that most people think that that is the case, and thus a good linen shift, its lack being a demonstrable sign of one's inferiority, carelessness and lack of thrift, becomes a necessity for the labourer rather than an indulgence. Smith recognizes that social judgements of morality can impose hardship and constraints on the poor even when that judgement is not rooted in a correct grasp of an underlying causal pattern. Deploying a Smithian lens to global development would therefore mean coming to grips with both the ambiguity of wants and the ambiguity of outcomesa bifocal approach that tracks the particular advantages and limits of the RCT.
The methodological strength of experimental randomized design is the ability to gauge the outcomes of a specific intervention. RCTs are a type of 'close epistemology', an approach that prioritizes realism of its representations rather than the elegance of its models (Kelly, 2009). The value of RCTs in development largely rests in their deflationary capacity to cast doubt on overly hyped claims of a new intervention's role in saving lives or easing people out of poverty. It was RCTs of microfinance initiatives, for example, that dealt a fatal blow to presumption that microfinance initiatives are playing a major or even a strong role in improving livelihoods globally. Banerjee, Duflo and others have spent over a decade studying microfinance programmes and began reporting as early as 2009 onwards that empirical evidence surrounding the efficacy of microfinance indicated modestly positive but not 'transformative' results (Banerjee et al., 2015, p. 1).
While this may seem like a benign rather than damning indictment of microfinance initiatives, the revelation that microfinance initiatives have led to only negligible gains for recipients has fuelled a tsunami of criticism over the amount of government aid and private philanthropy expended on rolling out microfinance programmes across the world. Governments and private donors continue to spend hundreds of millions each year subsidizing both non-profit and for-profit microfinance providers, offering 'aid' that fattens the revenue of microfinance providers even when the outcomes for loan recipients have been modest at best and harmful at worst, with many recipients experiencing crippling indebtedness (see Cull et al., 2016).
Criticism of government subsidies to for-profit microfinance lenders is just one example of growing discontent, voiced even by insiders within organizations such as the World Bank, with the use of development aid to subsidize commercial banking services (Gabor & Brooks, 2017;McGoey, 2015). Their concerns have not led to policy change at the World Bank or elsewhere, underscoring the fact that RCTs alone rarely produce sufficient evidence of a policy failure to divert government actors or private donors from pursuing economic policies that might fail to meet a stated objective, but which still serve their underlying interests or goals. In such instances, even a harmful policy can be rationally maintained despite the fact that a stated beneficiary loses out, bolstered by the efforts of actors who often find creative ways to strategically ignore inconvenient evidence (High et al., 2012;McGoey, 2007). This is the great 'failure' of RCTs: the way that even their clearest 'triumphs', such as the ability to furnish proof that microfinance gains are negligible, does not necessarily lead to policy change. RCTs adherents tend to presume that the revelation of policy 'failure' is sufficient to lead to policy reversal, but often this is not the case, and importantly the RCT method is incapable of providing a useful theory of inaction, because it is incapable of investigating the motives of policy actors who have unstated reasons for avoiding uncomfortable facts. In our final section below, we expand upon this point, juxtaposing our discussion with a summary of the six papers that make up this special issue.

'Leverage epistemologies' and their limits
The evidentiary promise of 'close epistemology' lies in its real-world empirical investigation in situ. But equally if not more compelling in leading to policy or at least ideational change are methodologies that harness the power of large data sets. We call these 'leverage epistemologies', because they have either computational, epidemiological or temporal qualities that allow for the exploitation of data on a potentially infinite scale. The clearest examples are new digital technologies that promise to unleash a new 'science of society' (Marres, 2017) with the same optimism that Montesquieu, Burke and Tocqueville once extended to empirical innovations in the late Enlightenment period. A recent type of 'leverage epistemology' is the historical and longitudinal analysis that Thomas Piketty applied to his investigation of inequality levels in a select number of developed countries over a period of more than 100 years, using a range of sources, from census data to tax records (McGoey, 2017;Piketty, 2014). In Piketty's case, it is the size of his data sets that lends moral authority and political impetus to the evidentiary claims made.
The example of Piketty is worth juxtaposing next to Duflo and Banerjee because, importantly, both made very similar claims about the value and the necessity of narrowing wealth disparities, but they adopted almost diametrically opposed methodologies for 'demonstrating' that there is sufficient evidence for government and private foundations to change their practices. Piketty has been outspoken about the limits of RCTs for studying large-scale macroeconomic change. As he suggested in a 2014 interview, 'We can't run a controlled experiment across the twentieth century'. He is still adamant, however, that our 'shared history' can lead to a robust understanding of the root causes of escalating inequality. In his words: 'All we have is our common historical experience, but I think this is enough to reach a number of fairly strong conclusions '. 4 Unfortunately, it is easy for his detractors to deride this evidentiary promise as Pollyannaish. Even among people who do broadly support and share his anxiety about growing wealth disparity there is endless disputation about how to mitigate the problem. As McGoey and Thiel point out in their contribution to this special issue, one of Piketty's critics is Bill Gates, who has stated that although he agrees with Piketty that wealth inequality is a problem, he disagrees on how to fix the problem. In particular, Gates insists that both 'philanthropy' and 'investing in companies' can and should offset a wealthy individual's obligation to pay more tax. Gates's belief that offering non-repayable grants to large corporations such as Mastercard can help to narrow economic inequalities is, however, unsubstantiated by evidence of any kind. Notably, it clashes with Piketty's primary insight: that spectacular financial returns to private investors over the past 60 years are outpacing national growth levels at an alarming pace, hindering the state's capacity to meet social needs.
At a time of sky-high corporate profitability, excessive CEO pay, the shifting of resources from R&D investment into marketing and advertising outlays, corporate tax evasion and the avoidance of liability for environmental harms, the idea that private corporations deserve non-repayable grants from tax-subsidized philanthropic foundations was until recently regarded as derisoryeven by the staunchest market fundamentalists on the libertarian right. The economist Milton Friedman, for example, scorned the notion that corporate actors deserve subsidies to execute a 'social' mission. In his words, if corporations are to 'make expenditures to foster "social" objectives, then political machinery must be set up to make the assessment of taxes and to determine through a political process the objectives to be served' (Friedman, 1970). Friedman was nervous that expanding the remit of the corporation would entail a necessary democratization of its operations.
Today, the corporation's remit has been expanded, but without a concomitant increase of democratic oversight or accountability. McGoey and Thiel's paper, as well as the contribution from Ehrenstein and Neyland, are focused on the ramifications of this development, contributing to a growing body of literature challenging the 'value for money' assumptions that organizations like the Gates Foundation draw on to rationalize their corporate enrichment gifting practices. Véra Ehrenstein and Daniel Neyland's paper explores the establishment of GAVI, an organization created in 2000 as a novel partnership between the World Health Organization, the World Bank, the Bill and Melinda Gates Foundation and UNICEF. GAVI claims to have introduced innovative ways to finance the purchase of vaccines developed by pharmaceutical companies, improving vaccination coverage across the world. Their paper is focused on a pneumococcal vaccine rolled out by GAVI over the last decade through the use of Advanced Market Commitment (AMC), GAVI's flagship initiative, billed as a novel 'market-based' mechanism to improve the research, development and delivery of interventions to treat neglected afflictions.
The demands of engaging in what Ehrenstein and Neyland call 'referential work', including the onus to demonstrate that scalability is feasible, sometimes necessitates, in their words, 'specific (and potentially questionable) evidential exigencies regarding disease rates, treatment efficacy, human bodies and germs, which are enmeshed with other considerations like political support and cost'. Ambiguities over the scale of the burden of the health risk and the cost of the vaccine pervaded the roll-out of the pneumococcal vaccine. One epidemiologist they interviewed stated that the harnessing of an early study showing considerable disease burden in order to advocate for more donor resources seemed a 'bit crude', given the cautious nature of preliminary results from the study, but the individual also admitted that advocacy helped to generate awareness about pneumococcal diseases and the need for vaccination.
A key pillar of pneumococcal advocacy work has been the reinvigorated emphasis on the notion, championed strenuously by the Gates Foundation, that every child's life has equal value. The insistence that every life is worthy of an equal chance at survival and flourishment has been an influential principle at the Foundation, leading Bill Gates personally to react angrily when people, such as Dambisa Moyo, the right-wing author of Dead aid, criticize the effectiveness of aid programmes. Gates referred to Moya's suggestions as 'evil', casting her criticism in a theological binary between the good and the malevolent (McGoey, 2015).
As Ehrenstein and Neyland's paper illustrates, however, stark declarations of metaphysical rights or wrongs tend to be a last resort for most global health actors, which instead rely on the ability to appropriate the evidentiary tools of an oppositional entity to raise questions of social welfare and health impact in the opposing party's own language. In an example reminiscent of Steven Epstein's (1996) study of the ways that HIV activists harnessed medical evidence to demand antiretroviral therapy, Ehrenstein and Neyland show how staff at Médecins Sans Frontières (MSF) adopted the economic language of costs and benefits to insist that the government subsidies used by GAVI to fund the vaccine were 'scandalously expensive', unjustly subsidizing corporations that had earned $19 billion from the provision of a vaccine that could have been developed more cheaply outside the AMC model. The opacity of commercial pricing strategiesdeemed protected information under confidentiality lawsposes a significant hurdle to judging the cost-effectiveness of new public-private partnerships like GAVI (see also Gabor & Brooks, 2017;Hickel, 2017).
Whether or not Western governments overpaid private pharmaceutical companies to deliver the pneumococcal vaccine is not simply an academic question. Resource availability adheres to a zero-sum logic: the more that private companies gain from subsidies, the more slashes in expenditure a state actor may need to make elsewhere. As scholars such as Macklin (2014) point out, all governments have a responsibility to take appropriate steps to fulfil the human right to health, described in Article 12 of the International Covenant on Economic, Social and Cultural Rights as 'the right of everyone to the enjoyment of the highest attainable standard of physical and mental health' (Macklin, 2014;Piccard, 2011). But the US government is notoriously reluctant to support human rights treaties, and the Covenant on Economic, Social and Cultural Rights is no exception: the United States has signed but not ratified it, making it one of a half-dozen of UN member states to do so. Growing poverty and worsening health outcomes in wealthy nations such as the United States raise questions surrounding the morality of tax deductions for large philanthropic organizations which use their tax-exempt privileges to offer non-repayable grants to private pharmaceutical companies seen by Macklin and others as charging exorbitant prices domestically.
The question of whether increasingly influential philanthrocapitalist organizations like the Gates Foundation and Bloomberg Philanthropies are sufficiently accountable to the public is the focus of David Reubi's paper. Drawing together insights from scholarship on audit cultures and extensive empirical research on the global tobacco control campaign spearheaded by the Bloomberg Initiative, Reubi challenges the idea that private foundations rarely respond to, let alone act upon external dissent. He shows how the respective investments of the Gates and Bloomberg Foundations in the production of public health metrics, project reports, log-sheets and data tracking operate as a form of accountability in itself, answering the call to bring more transparency to decision-making at private foundations. Comparing the performative powers of this epidemiological-meets-managerial quantitative deluge, Reubi raises under-examined questions about the limits and nature of democratic accountability in global health. In some ways, the Gates Foundation's willingness to admit mistakes is itself a form of evidential 'referential work', in Ehrenstein and Neyland's terms, signalling its open-mindedness and thus its legitimacy to influence government spending priorities. By admitting its own earlier failures, the Foundation garners legitimacy while deflecting larger questions over what legal obligations non-democratic organizations that 'affect to trade for the public good' owe to the public. By bringing Bloomberg and Gates into comparative light, Reubi illuminates that the nature of those obligations also hinge upon which publicshareholders, donors, constituents, citizens or an epidemiological body politicis being taken into account (cf. Krause, 2014;Mahajan, 2018).
Reubi's paper opens with a quip from Bloomberg -'in God we trust; everyone else bring data'. The ecstatic power of data to garner faith and mobilize action in the face of uncertainty forms the focus of Ann H. Kelly's paper. Her work with the World Health Organization (WHO) during the 2013-2016 Ebola outbreak provides the prompt to interrogate the distinct styles of reasoning that usher the transformation of a public health crisis into a global security concern and configure a humanitarian emergency into a site of pharmaceutical investment. Kelly directs her attention to the accelerated development, testing and licensure of Ebola vaccinesan experimental and potentially 'game-changing' intervention that captured the imagination of the global health community. Introducing an unproven and, indeed, save for animal studies, barely tested vaccine into an emergency raised a host of ethical, logistical, financial and political problems. Moreover, the epidemiological justifications for testing the vaccine were built upon an absence of data, a fear of what was missing from national reports of incidence, rather than on deductions from the information available. Kelly shows how the modes of uncertainty attendant to the declaration of the Ebola outbreak as a public health emergency of international concern (PHEIC) created the context in which suspension of thoroughgoing empiricism was a necessary and, ultimately, moral duty. 'The imminence and inscrutability of risk', Kelly argues, 'created the space for political action' (2018).
The forms of fact-making that emerge under these conditions of public health exigency and epistemic deficit betray a heady blend of moral logic, economic interest and prophetic vision. Following the debates over the design of clinical trials as the vaccines made their way from the lab into the field, Kelly shows how the balance between humanitarian needs and scientific rigour was struck through an appeal to history: the WHO's triumphant smallpox eradication programme provided the inspiration for the 'ring vaccination trial' that was successfully implemented in Guinea. As new standards of research practice are being formulated for an emerging paradigm of emergency research and development, Kelly believes that the experience during the Ebola outbreak generates critical lessons into the forms of legitimacy that drive collective and institutional action in times of public health crisis. Kevin Donovan's paper provides a rich context to those new norms, exploring the uncertainties over which methods produce the most 'rigorous' evidence of both the causality and the effectiveness of various interventions. He focuses on RCTs in development, tracing the epistemic affinities and thought collectives central to the growing dominance of 'randomistas' in public health. One of his central points is that advocates of RCTs in development often switch registers when propagating the value of RCTs to different audiences. When publishing findings in economics journals, practitioners are careful to ensure that any causal claims are robustly defended; when communicating with policymakers, Donovan notes, 'the indeterminacies of accountability and cost-benefit analysis are emphasized'; and when addressing lay audiences, 'uncertainty is to be reduced because it impedes the effectiveness of aid and, therefore, contributes to human suffering'. Donovan notes that the flexible ability to range between rhetorical registers, alternately stressing and trivializing the problem of uncertainty, helps RCT advocates 'to enrol numerous supporters'.
The capacity for different social actors to gauge and choose which 'face' to assert publicly is, of course, an old observation of the social sciences, but the role that empirically oriented economists play in adding moral gravitas to funding decisions remains under-theorized. Like Reubi, Donovan's paper makes new and original links between older work by Michael Power and others on 'audit culture' and the problem of policy circularity within development arenas. As he suggests: 'it is a tacking back and forthbetween promises of certainty and accusations of uncertainty; between registers of argumentation; and between audiencesthat has helped reconstitute the spirit of international aid … this lack of certitude has not called into question certainty writ large but served to justify further experimentation'. Véra Ehrenstein's analysis of UN negotiations on global warming demonstrates the complexity of those registers of argumentation when the object of intervention is in doubt. Ehrenstein begins where most social science studies of the politics of global environmental science end. For the political actors that form the focus of her study, the evidence for climate change is not in dispute. Yet, consensus on the importance of reducing global carbon emission does not in any way suggest a clear way forward. Drawing from ethnographic work with UN decision-making bodies and, in particular, the efforts to protect rainforests in the Democratic Republic of Congo, Ehrenstein demonstrates how the art of negotiation becomes an end in and of itself, a means to performatively address divergent concerns while neutralizing the potential of public discord. Like Kelly, Ehrenstein points to the critical role an imagined future plays in creating a sense of shared purposepredictions and promissory commitments of financial support for forest protection laws and afforestation activity can help occlude the challenging pragmatics of implementation, and the further, thornier question of how to demonstrate their impact. Indeed, Ehrenstein shows how the evidentiary demands of these policies remain underdetermined, a metrological vagueness that on the face of it speaks of a commitment to inclusion of a diversity of metrics, but also betrays an 'optimistic unrealism' that might never yield the futures these policies hope to sustain. In other words, uncertainty over the success and failure of past development gambits demands resolution, which powerful actors, by virtue of their perceived expertise, level of financial resources or institutional obligation, are enlisted to fulfil.
Increasingly, actors like the Gates Foundation and Bloomberg Philanthropies uphold themselves as 'partners' in the provision of public services, despite the fact that the question of what legitimates and obligates them to serve the public remains unanswered. Until now, the 'charismatic authority' wielded by powerful philanthropists has largely been seen as sufficient to legitimate their growing influence over human affairs. This authority comes in part from the ability of private philanthropists to selectively publicize global health 'miracles' even while troubling questions over cost-effectiveness and corporate harm to the public are ignored. Indeed, the tangled vested interests of these actors with the pharmaceutical, tobacco, food and drink industries tend to inflate the value of market-based initiatives while obscuring the root causes of collective distress (see Herrick, 2009Herrick, , 2017.
Thus, in an iterative fashion, today's increasingly powerful donor class of wealth elites is able to harness the power of both the 'new empiricism' and what we have called the 'new theology', by moving flexibly between empirical defences of private influence over public spending, on the one hand, and semi-religious defences of the same, on the other. This tacking process is never entirely divorced from secular pressures to defend one's evidence base through rational rather than spiritual rationales. Indeed, this has been a fundamental goal of the Gates Foundation: to use data to show that aid 'works' in saving lives. But when the Foundation's critics use a similar methodological approach focused to show that aid initiatives led to meagre or damning results, as Moyo's work has suggested, the Foundation is able to rely on its earlier, principled approach to givingthe belief that 'every life has equal value'to dismiss criticism as amoral or sacrilegious. A further way to conceptualize this flexibility is to suggest that private foundations have the freedom to shift easily from the framework of a biopoliticsstressing the capacity of private gifts to emulate the gift of life, literally 'saving' souls through private actionto a framework of necropolitics, insisting that more children will die if aid or philanthropy is ever criticized. The larger question of whether private actors deserve to play a determinant role in choosing who lives and who dies is strangely elided from debate even as key players harness a language of life and a language of death to defend their influence and power.

Conclusion
As the papers in this special issue make clear, questions over which actors are best placed to deliver public services lie at the centre of the democratic social contract. In recent years, this debate has increasingly been framed in a narrow way, in a language that favours discussions of efficiency over equity, costs and benefits over the right to access. But an emphasis on 'what works' cannot be disentangled from the question of 'who counts'. It is true, for instance, that vaccine coverage has improved over the past 20 years across developing regions. It is also true that the cost has risen enormously, placing pressure on developing and wealthy countries to rescind availability or divert resources from other health initiatives, leading to both saved lives but also to an increase in unnecessary and preventable childhood deaths (Macklin, 2014).
Questions over 'who counts' are, of course, old ones. The eighteenth-century statesman Edmund Burke was simply one of many late Enlightenment figures to offer a spirited defence of the right of aristocratic classes in Europe to impart charity on their own terms, free of the forced obligation to fulfil the 'right' to food or shelter. He drew on both the 'new empiricism' and the 'old theology' of his day to make his points. In the same vein as increasingly anti-egalitarian Western authorities such as the US government today, Burke insisted that having the right personnel devoted to the task of determining optimal resource allocation is more important than any abstract claim to a 'right' to social and economic security. But this draws us back into Smithian territory, and the problem of false perceptions of morality and liability for poverty. Burke is wrong to suggest that the 'right to food or medicine' is irrelevant to resource allocation, because recognition of a citizen's democratic right to challenge state decision-making is different from the obligation in non-democratic states to submit to a ruler's diktats.
This is a point that is well-traversed in political theory, but the epistemological problem with Burke's statement may be less apparent. Burke was a contemporary of Smith's. It is likely that many, if not most, of Burke's 'farmers' and 'physicians' whose expertise he appeals to would have shared the presumption about the poor that Smith warned us about: the tendency to blame the poor for their destitution, colouring the ability to adjudicate fairly or 'truthfully' over what the poor deserve, because Burke's own cultural biases limit the ability to determine the 'truth'. Similar false presumptions about where the responsibility for poverty lies are underpinning contemporary austerity measures both domestically and at a geopolitical level. While 'populist' and 'liberal' Western leaders alike may express a commitment to foreign aid or philanthropic bequests, reparations for recent neocolonial acts, such as the invasion of Iraq, or for more longstanding processes of colonial subjugation seem to fall outside the scope of humanitarian responsibility and remain an anathema to mainstream policy discussion.
Today, particularly in countries that have ratified the Covenant on Economic, Social and Cultural Rights, citizens have a legal right to challenge the lack of adequate health provision. In the United States, the ability to point out the US derogation of responsibilities can help to attach 'shame and blame to governmental neglect' (Yamin, 2005). But an emphasis on neglect is less compelling when it comes to private actors. Any effort to 'shame' a private philanthropic entity for its voluntary health expenditures leads to blank stares or even anger, as people query the morality of attacking charityeven when philanthropic disbursements can exacerbate the food shortages, ecological strife and wealth disparities that fuel the need for philanthropic bequests in the first place. An increasingly powerful group of global policymakers today often appeals to the robustness of empirical evidence, such as the lives saved through vaccines, to justify resource allocations that leave the political economy of vaccines production unchallenged. In short, the denunciation of the 'right to food and medicine' that Burke disparaged at the onset of the French Revolution is as powerful in our time as it was in Burke's.
From the effort to contain the spread of infectious disease to the mushrooming of chronic illness across developing and rich countries, the framing of the world's most pressing problems hinges on whether a crisis is seen as sufficiently urgent; on the question of whose death 'counts' the most. Collectively, the papers in this special issue raise new questions about the relationship between evidence and resource allocation; between empiricism and 'enlightened' decision-making. By returning to conflicts that pervaded the late Enlightenment period, our introduction has emphasized a basic but often forgotten point: whether a child, man or woman 'deserves' to live or die often depends on whether they are seen as an equal human being. The entitlement to stratify and to categorize social needs cannot be said to be 'enlightened' if a receiving demographic lacks the power to question the imposition of economic policies that 'help' them in ways that cannot be rejected or overturned. Should the right to challenge global resource allocation be the purview of the few, or the right of the many? Whether one sides with Burke or Wollstonecraft is not a moot point. Throughout the modern period, the acceptance of new truths, of more truth, has rarely been a peaceful enterprise. Rather, the willingness to concede that enslaved or conquered peoples 'should be accorded membership of the family of mankind' has been achieved only through the militant righting of 'truths' upheld by good statesmen as godly gospel. In the meantime, the feudal hoarding of finite resources continues, unmolested by the demands of populist masses who only in time are lionized as fervently as they were once condemned.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
The Notes 1 We borrow the phrase 'promiscuous pervasiveness' from Neil Brenner and colleagues, who use it in a different context to describe neoliberalism. 2 Thanks to Michael Halewood for the reference to Gilroy. 3 Most scholarship tends to exclusively focus on the primary meaning conveyed by Smith, but some scholarship has attended to the second meaning. See in particular Davis and Sanchez-Martinez (2014). See also recent studies which explore Smith's complex, counter-intuitive and neglected writing on the problem and root causes of wealth inequality (Boucoyannis, 2013;Rasmussen, 2016). 4 http://blogs.lse.ac.uk/politicsandpolicy/five-minutes-with-thomas-piketty-wedont-need-19th-century-style-inequality-to-generate-growth-in-the-21st-century/.