This contribution focuses on the use of quantification in the new governance techniques that emerged for the most part in the 1980s, first in the United States and Britain, before spreading to Europe and the rest of the world under the auspices of international organizations such as the World Bank, OECD or the European Union (OECD, 1994). In this use, “quantification” refers to maximizing quantitative objectives to be achieved through definition, implementation and supervision of policies, either by management rules in organizations and companies, or by measures adopted in the context of public policy. Such techniques have profoundly altered the practice and final purpose of quantification. Far from underpinning statistical observation of reality, quantification is now expected to serve political measures that are proposed, or already decided. Quantification is expected, not only to test them, but is also enjoined to demonstrate their efficiency over time or in comparison with other policies. Its master concept is performance, which witnesses an inversion of priority in the use of data in politics: not only aimed at measuring, but, above all, performing. The historical connection of statistics with implementing public policies (as described by Desrosières, 1998 [1993]; Hacking, 1990) is taken over by performance indicators that both orient and evaluate outcomes, objective by objective.
Governance seeks to place effectiveness at the core of collective action, whether in organizations or in government administrations. This effectiveness is measured by performance indicators. These indicators pertain, directly or indirectly, at resources and means granted, to results expected from better management of the company or the administration. These reforms are justified by the stated (and debatable) claim that “more” is the equivalent of “better”. It is the role of quantification to show that this is indeed the case, by internalizing this definition as a key step in reform.
Anzeige
The mainstream economy wholeheartedly applauds this targeted focus on efficiency. It took time for practitioners of political science to become aware of the issue, at least partially, especially for those not interested or not familiar with data construction’s subtleties.1 The sociology of quantification (in particular the branch derived from the sociology of science) sees in this phenomenon a field of research related to its habitual domain, with description of these new quantification instruments and practices, particularly within states, and according to the nature of these states. In this book, and elsewhere, one finds significant contributions to the sociology of quantification in these areas.
Introduction: Towards Governance-Driven Quantification
There is nonetheless something different, something more than the focus on effectiveness; there is a sort of “revolution” of quantification, when it is governance-driven. Three aspects emerge, which, we will see, challenge democracy as a government procedure and collective practice.
First, the objectivity of figures is used as a political argument. Figures do not lie. In itself a figure tells the truth of the moment on the question at hand. This truth is of course approximate, because truth is beyond our reach in this world. But this approximation is taken to vouch for the seriousness of the figure, and of the arguments based on the figure. Debate is developing mostly on the existence and amplitude of the margin of error, and not on the political relevance of the figure produced (that resides in the details of their modes of definition and calculation).
Second, the qualitative complexity and diversity of social and historical processes are reduced to a few quantitative scales of appraisal. The notions of equivalence and comparability, and a static perception, are introduced as tools of analysis, whereas many phenomena are singular, not commensurable to others, and are part of a dynamic group process.
Anzeige
Last but not least, the rational construction of data by the implanted quantification processes tends to be formatted for proving that the policy implemented is appropriate and successful. This capacity of self-producing the politically expected data is one of the most significant innovations of governance techniques, one that is surprising and hard to understand for the non-specialists.
To apprehend this development we will take a long view of the social history of quantification by looking at a specific example, that of its role in the emergence and decline of unemployment as a social category. For the effects of governance are not limited to quantification instruments. By capillarity this governance—as an emerging phenomenon, rather than as a rational project—gives birth to another political, social, financial and economic world that alters the way governance sees itself, how it frames and analyses problems, how it acts and evaluates its action.
By comparing the role of quantification (when it was known as statistics) during the “invention” of the unemployment category, and then during its decline, we can see what is different and what is new in today’s governance-driven quantification. The core of the changing between the two is the status and the role given to democracy and participation of people, both in political and quantification processes.
In brief, the purpose of statistics is to build general knowledge “extracted” from the plurality and variety of social conventions people use in daily life to understand their world, to coordinate with others, to pursue their aims and try to achieve their ends; and on its knowledge basis to define policies apt to meet these conventions (see also Desrosières, 2011). By fabricating cognitive proofs that things are going the right normative way, governance-driven quantification becomes part of the political process: producing knowledge becomes the oriented by-product of politics. The purpose of governance-driven quantification is to find ways to rationally transform social conventions towards some pre-given political objective, judged by the Centre as optimal.2 So in this case, as we will see below, such quantification could be best defined as “inverted statistics”. In my view, the what-works approach that led to evidence-based policies (see in particular Davies et al., 2000) has been the premise of such an orientation which, afterwards, has been developed into technologies of management by performance.
The history of unemployment as a social category comprises two periods, its rise, and its fall. Unemployment emerged in Europe, roughly from 1880 to the 1950s, as a social category to be elaborated, measured and targeted by public policies. At the same time, what was then called statistical science, collective reflection and thinking about the instruments and uses of statistics, emerged. After a period of stability, through the 1980s, this social category began to decline. For Europe, the progressive relegation and probable future disappearance of this category as a public social concern are the paradoxical fruit of the European Employment Strategy (EES). It is not that this disappearance is deliberately intended, quite the contrary, the EES aims to increase employment levels, by raising the employment rate in the population of working age. Along with other factors (and the evolution of the labour market) it is the result of a choice made by the European Union in the 1990s, to monitor employment policy using performance-based governance techniques. As employment policy remains the prerogative of states, the European Commission invented a system of voluntary coordination of national policies, called the Open Method of Coordination (OMC). It has been extended at the European level to other social domains: social inclusion, pensions, health/long-term care, and to the Broad Economic Policy Guidelines (BEPG). The launching period 1997–2006 was crucial in that the European institutions boosted collective learning by national senior civil servants of the method and, more generally combined with other influences, of New Public Management methods.
In the next section (Part I), we review how far producing and interpreting data rely upon institutional machineries and, often neglected, upon the participation of inquired people. In the subsequent section (Part II), lessons that can be drawn from the socio-historical invention and deconstruction of the category “unemployment” for three European countries, France, Germany and the UK, will be discussed. The turn from statistics to governance-driven quantification is illustrated by the way European institutions deconstructed the category. The following section (Part III) draws the implications for democracy from the turn towards governance-driven quantification. It emphasizes the political move towards “a-democracy”. The final section (Part IV) explores ways by which social criticism can oppose this turn by taking on board justice expectations into quantification processes and, in so doing, make way for reintroducing democracy. To be just quantification must be correct and fair is the message implicitly sent by Amartya Sen when he puts forward his concept of informational basis of judgement in justice.
Producing and Interpreting Data is a Collective Undertaking
Most often, if not always, quantitative data are taken at prima facie by users. Data present themselves as evidence. For users data are “real”, or tend towards a pure reflection of this “real”. Thanks to them collective decisions are evaluated, undertaken and followed. Such beliefs neglect the fact that data are produced and interpreted along a chain of several steps, in specific configurations of actors in which statisticians or quantifiers are involved with other actors, in deliberative arenas. Any data process should be viewed from two sides: the institutional machinery organizing the process on one side; and, on the other side, (very often neglected or even forgotten) the people who, through their answers, are the object and support of the searched data.
The institutional machinery could be directly that of the state, or that of a firm or any collective organization requiring data. In this second case, the state is indirectly present through public regulation and law. The main components of the machinery are: the conception along which the state is built; the questioning and its tools (organization of the questionnaire; the type of inquiry and its methodology or administrative requirements in case of data as by-products of administrations or management services); the instructions for coding the answers; the production of statistical tables (which requires nomenclatures, categories to classify answers and rules to aggregate individual answers in order to put every person into one case and only one). All these components play their role along a chain of production with many steps; each of them open to several possible technical options and to different interpretations; all managed by sets of organizational rules.
As outcomes of this chain, users have at their disposal a wide statistical material: variables, tables, correlations between variables, dispersion figures, indicators, and so on. They can quietly assert, for instance, that “the number of unemployed people is this”, “the rate of unemployment is that”. In so doing, they neglect that the data they use have been produced along a chain of production in which many not neutral technical choices are to be made. They also neglect the second side of data, namely that their primary resource, like coal or iron, has been worked out, is constituted of persons who have to answer or to be classified.
The beliefs about questioned people oscillate between two extremes. In one extreme, they have no margin, except to provide the expected right answer. They are viewed as passive resources or sites automatically responding to some external stimulus, like in behavioural models of experience. At the other extreme, they are viewed as pure rational cheaters who have to be severely controlled. These are both dire mistakes against which quite nobody (be it politicians, technocrats, civil servants, economists and statisticians) can be taken, at diverse degrees, as protected.
Firstly, such beliefs impede us to see what one could call “the democratic paradox” in our (until now) democratic societies. To understand such a paradox, it is necessary to have, in data production, a wider view of “democratization” than usual in politics. The first step is to be aware that asking people to respond to questionnaires basically means that, “somewhere”, their answer has some intrinsic value and should be collected as such. Not as pure and transparent carriers of some pre-existing underlying reality, the standard view, but as active interpreters bringing some practical experience and knowledge of enough value to be used in collective choices. When collecting their world views and experience of the domain at stake, persons become active mediators and go-between between the supposed real and the data. Their experience has to be considered as having a knowledge value. The second step is to take into account what they have to say on them when defining categories and methodologies.
Secondly, data are built upon the “official understanding” of the investigated domain.3 However, depending on their situations of life and work, their biography and life course, people have varied experiences with regards to this domain. There are many personal or collective understandings of the same reality, each being a priori as effective and relevant as the others. Sometimes the major part of these understandings could differ from the “official understanding” which forms the basis of the questioning. Basically, the intrinsic value of individual answers does not depend on their good will to answer, or on their correctly answering in the sense of adhering to the official meaning. This value is elsewhere, in its potentiality to reveal distances between different understandings for the same “object”, which leads us to the third point.
Thirdly, and not the least, such intrinsic value is in essence democratic. For it has the capacity to put the spotlight on the distance (and to open a window on its meaning) between the understanding a person has of her situation and the questioning incorporated in the questionnaire.4 These distances or gaps between individuals and official understandings on the same domain signal the existence of the plurality of possible relevant questionings (Boltanski & Thévenot, 1983; Thévenot, 1983). The official one is one among others. No data, especially aggregated data, can be said to be the truth, not only because they are deeply linked to the series of both technical and political choices made along their chain of production, but basically because among a range of possible choices, one path only has been chosen.
One will see the huge impact of all these factors on the nature of data produced in the three countries we will review below: France, Germany and the UK.
Inventing and Deconstructing Unemployment as a Category: The Role of Quantification
Almost at the same historical period (the turn of the twentieth century), unemployment as a social category and as a procedure to count those to be classified as unemployed was invented in the major European countries: France, Germany and the UK. Such inventions lasted half a century or more. The national processes and their outcomes were very deeply anchored into national specificities. They brought to people and their political communities new resources to understand “their” real, to act within it, to form expectations and projects, to legitimate decisions, disagreements and conflicts. Invention has followed the road of statistics as we suggested in the introduction. Statisticians were involved in diverse deliberative arenas, and were at the initiative to create both the category and the methodology. Deconstruction is following the road of governance-driven quantification (for a detailed demonstration see Salais, 2007). It disqualifies ancient and familiar resources that offered stable anchors for people, without, until now, providing alternative types of resources.
The Invention of Unemployment: Comparing France, Germany and the UK
The “invention” of the category unemployment at the turn of the twentieth century demonstrates how far data (categories, procedures, numbers) are worked and re- worked all along the process by the actors, in a sort of joint production.5 There was a kind of double plurality at work, the plurality of institutional machineries among national states and within them, on one side, and the plurality of indigenous categories among people on the other side. In each country, the data to be produced had to meet a demand for information which was linked not only to public policies but, above all, to their specific conceptions of collective objectives and to the way the state should intervene for their achievement. Formats, specifications, levels of collecting and using data, even the need to collect or not, all will depend—for the same domain of observation—on these state specificities, which were, and still are, very diverse among countries and over time. States were more or less inclined to systematic and general quantification, more or less open to democratization of data production. They required different types of data and of their “production system”. From the people’s side, there was another type of plurality, one of the principles of justice considered as legitimate on which to build the category.6 In the final two sections (Parts III and IV) we will connect democratization and justice.
In France, whose state has been historically built along top-down, systematic and central intervention through general categories, the search for defining unemployment has been undertaken directly by elites surrounding the central state administration. Lawyers and economists (at that time trained into the same faculties), statisticians, economic and social actors, members of the parliament, public officers, tried to have their word, using their own knowledge and experience. They met in different assemblies, circles and savant societies (Didry, 2002). In the 1890s, the state created a special institution, named “Office du travail” which launched inquiries, monographs, collected professional advices to have a clear understanding of the various work conventions especially with regards to periods of no work (Luciani, 1992). All together were able to define a general and practicable category of unemployment which was incorporated into the census and administered to the whole population, for the first time, in 1896. All French administrative levels were progressively required to use the same category and to produce the same types of statistics and tables at all administrative levels. However, the disparities in the rates of unemployment among regions, professions or labour statuses reveal durable traces of other conceptions, especially homeworkers and independent workers, employees of local small firms, craft workers who have their own conception of the primacy of individual, local or craft responsibility in social compensation or in job search organization (see Salais et al., 1986, chapter 3).
In Germany, a federal state, such a national unification of the category failed. There were already a series of local definitions and conventions, depending on the professions, the unions and the towns. These definitions were founded on specific principles of justice that led to various principles to identify unemployed people: for instance, belonging to local crafts; being citizen of the local town; being registered on local social help bureaus (Zimmermann, 2001, pp. 126–138). Land statisticians, convoked to Berlin, were unable to agree to a common definition. It is only in 1927 that some unification was achieved, thanks to the national social insurance system which was eager to generalize insurance to unemployed situations. Yet, being centred on previous craft insurance systems, it tended to exclude workers that did not belong to craft unions. A categorization of the unemployed appeared for the first time only in the 1931 population census. This was not renewed by the Nazi regime in the 1936 census. Beyond the failure to generalize, it shows that, except for the Nazi period, Germany as a national entity is built along with a different conception of the state, mostly that of one we call a “situated” state,7 a concept we develop in Salais and Storper (1993, fourth part). Such a state gives precedence to collective autonomy over national top-down intervention. It tolerates diversity; the responsibility to define the common good at stake and to take care of it can be left to various levels, especially the Land, the city, the profession or the economic sector. Statistics can have different frameworks and tables for the same domain, which leaves some collective freedom to choose the relevant principle of justice for building the data.
In the UK, the historical picture was also another one, a long and uncertain battle between at least three conceptions of unemployment, implying the state only indirectly, and of the assignation for responsibility: poor laws, trade unions or the market. All these systems had their own statistical categories and data which were not consistent with each other. Poor law, the oldest system, was placed under the sovereignty of the King, but managed at the very local level of the parishes; unemployed people were not differentiated from the poor and treated as such. Trade unions had their own system for their members. Unemployed members were supported by friendly societies which did not differentiate between the lack of work due to unemployment or strike; both were financially helped (Phillips & Whiteside, 1985). At the turn of the twentieth century, social reformers (the most famous, among many, being William Beveridge) were hostile to these systems. They pleaded—with more or less success—in favour of the creation of labour offices which could rationally construct a true national labour market. Such a market not only should work as a perfect market, but have the tasks to clear the market from the unemployable (sent to other social policies) and to teach workers to be individually responsible for their situation and their future (Mansfield, 1992). Regarding unemployment, the UK has thus implemented contradictory conceptions of the state, valid at certain levels and for some organizations, but not at others: interventionist for constructing from the top the perfect market, but in competition with local autonomy and professional diversity which would have been best taken in charge by a “situated” state. Several principles of justice are in competition to define and observe unemployment, presumably in some unstable compromises even today, based on, respectively: the deserving poor, the acknowledgement by peers, the morally regular worker (Whiteside, 2014).
Governance-Driven Quantification as Inverted Statistics: Europe and the Reversal of the Pyramid
However, a new actor appears on the field of employment in the 1990s: Europe, its institutions and political frameworks (for a historical perspective on building Europe see Salais, 2013). It added complexity, more uncertainty in the definition and observation of unemployment, and in the meaning of data. Basically, it contributed to blurring the boundaries within established categorizations and to deconstructing them. Especially, short unworked periods are less and less considered as “unemployment”, but as transitions—that have to be the shortest possible—between two jobs or tasks. In practice for part of the population it corresponds to precariousness, but precariousness is not recognized as a valuable category of social policy and not counted as such (see Standing, 2014).
European institutions introduce new public management reforms through a specific method, called the open method of coordination (OMC). This method constitutes a fascinating illustration of the social and political impact of quantification when internalized into governance schemes. It reveals its basic specificities. We will pass in review five of them: the reversal of the statistical pyramid; a new target for employment policies; statistical tables as driving forces; the set of indicators as embedded norms and guidelines as justificatory covers; a cooperative game between rational actors.
The Reversal of the Statistical Pyramid
Governance-driven quantification operates a reversal of the pyramid which, in classic statistics, links its large basis (the multiplicity of individual experiences and the mobilization of their social knowledge of situations and problems) to its top (the producing of aggregated data, via the progressive reduction to numbers by aggregating individuals’ answers). Governance-driven quantification puts the pyramid not on its basis, but on its top. It starts from the top data (the quantitative global performance) to be maximized at all costs whatever the means used to achieve this objective. It tries, through a descending movement, to produce the required basis of the pyramid able to generate the expected global outcome. Quantification rules of measurement, organizational rules of political schemes are adjusted in order to fabricate, if not individual behaviours themselves, at least answers, or statistical treatments that fit with the quantification objectives. The underlying utopia of quantification and, as a consequence, of governance by numbers (see Miller & O’Leary, 1987; Miller, 1992; Supiot, 2015), is that social subjects are expected to create by themselves a reality that complies with the objectives. They would, eventually, spontaneously produce the required data. In general these are only answers that, through several organizational means, at the end begin to fit with maximizing the scores. Such utopia to make people spontaneously creating an “optimal” social reality must not be confused with the ordinary faking of statistical data, frequent on sensible domains like unemployment statistics.
A New Target for Employment Policies
In the European employment policies promoted since the end of 1990s, European Union authorities took the global rate of employment as one of its major macroeconomic indicators.9 They substituted the search of Keynesian full employment for the maximizing of the rate of employment as their main target. In so doing, a “job” is no longer what it promised to be in the model of full employment. In that model, any employment guarantees minimum standards of remuneration, of security in the face of unforeseeable events based upon social and economic rights. What Europe now guarantees to its citizens was only to have a task, whatever it could be and under the condition they accept it.
In practice, to measure the national rate of employment, the European authorities recommended applying the definition that is used by the ILO to build international statistics on employment: “Employed persons consist of those persons who during the reference week did any work for pay or profit for at least one hour, or were not working but had jobs from which they were temporarily absent” (see, for instance, European Commission, 2006). Statistically speaking, applying this definition is simply following the ILO definition.
But it takes on a very different meaning when it is translated into political action. It means that, whatever the task is in terms of quality (wage, working conditions, duration, type of labour contract), it can be considered as employment if it lasts at least one hour a week. All other characteristics were deemed irrelevant when creating employment data. One should call this “the convention of employment without quality”. This convention is far from trivial. Employment without quality is a task stripped of all legislative guarantees (in terms of recruitment, protection against unfair dismissal, minimum starting wage) and social provisions (social and economic rights). By removing quality features when comparing and putting in competition their social systems by means of such single quantitative scale, the Member States are encouraged to water down the quality of their employment conventions in order to improve their quantitative performance.
Statistical Tables as Driving Forces
One should pay attention to what is ordinarily taken for neutral, hence unproblematic, that is the collection of statistical tables that, for each yearly report, national administrations are required to fulfil in the areas using the OMC. One must suspect that, to a large extent, these tables are the driving forces “behind” the formalism, not only for data, but, beyond, for political discourse (vocabulary and syntax). Tables also act as rhetorical justifications of the normative background imbedded and for most people dissimulated in data, especially in the selected indicators.
Contrary to the standard view, a table is not only a collection of figures (one in each box, for instance, as in a double-entry table), some being higher and others lower, from which one can directly draw conclusions like “the female rate of employment in 2005 is higher in the UK than in France”. A table is, above all, a procedure for aggregating individual situations, for instance, relating to employment and the person’s position in the labour market as built by nomenclatures. All situations compiled in the table which are considered as identical with regards to these two nomenclatures are placed in one box. They are considered as equivalent according to the corresponding properties. In other terms, filling a table by combining individual data requires conventions of equivalence,10 which decide about what should be considered as similar. These conventions ensure the passage from the particular to the general (what Luc Boltanski & Laurent Thévenot (2006 [1991]) call the rising into generality).
Generally speaking, conventions of equivalence are ignored or misunderstood by the ordinary users. From the above statement on female rates, users will spontaneously conclude that “women work less in France than in the United Kingdom”. But this conclusion is valid only if the legal, statistical and social definitions of what should be considered as a “job” are identical in the two countries. In practice, the UK is using a “softer” definition of part-time work than France, which results in women who work very few hours a week being considered as having a job and driving them into such jobs. The situation is even reinforced with the invention in the UK of the zero hour contract. Applicants are asked to stay available at home for whatever task and at whatever moment their employer decides. They are considered as employed even with zero worked hours. This helps maximize the rate of employment in which the UK is champion (which, as a counterpart, corresponds to one of the highest poverty rates in Europe).
The Set of Indicators as Embedded Norms—Guidelines as Justificatory Covers
Conventions of equivalence govern what we select, what we exclude and what we construct. Thus, the requested description becomes not far removed from a normative evaluation of the situation under review.
The basic issue with the Open Method of Coordination—and more generally governance-driven quantification—is not immediate strategic action, or neoliberal ideology11; it is about the cognitive conventions that are selected to drive the political process. The selected set of indicators frames the normative background of the political decision-making process. It is neither malignity nor political cunning. It is the mere consequence of the fact that any indicator selects what is worth to be known or not and, in so doing, basically builds the reality that is relevant both for the deliberative process preceding the decision and for the action to be undertaken.
The set of monitoring indicators selected by the European Employment Strategy (EES) focuses on the supply side of the labour market, which is the work offer by the manpower. It expresses the norm that work offer should be the highest, the most flexible and adaptable to economic hazards as possible. Employability is the main concept. The higher it is for an individual, the more he would have access to job opportunities. At first glance, there is no problem here. But complete labour market models emphasize a second concept at the same level of relevance, the one of vulnerability to job losses. The more you are vulnerable to job loss, the less you could access a stable job. So employability should go hand in hand with job security (or at least stability) as objectives for employment policies. There is nothing like this in the EES. Furthermore, the monitoring indicator for evaluating employability is the rate of return to employment. The fastest it is, the best it is for the EES. But improving employability is wider than increasing performance, for it has qualitative aspects that, normally speaking, should be taken on board by public policies, but are not.
Here appears the mismatch between the political rhetorical justification one can see in the wording of guidelines, and the effective policies that are driven by their monitoring of performance indicators. The search for consistency between data and discourse is, in effect, part of the global drift from politics to management. It tries to be achieved through the connection between quantitative monitoring indicators and guidelines that are expressing the objectives corresponding to the different indicators. There is a rather subtle, but essential shift of normative requirements from guidelines to indicators. The formal normativity is provided by the guideline, the effective normativity by the indicators to maximize. One will take the example of the EES guideline “Ensure inclusive labour markets”, introduced in 2006 (European Commission, 2006). It asks the Member States to develop “active and preventive measures including early identification of needs, job search assistance, guidance and training as part of personalized action plans, provision of necessary social services to support the inclusion of the furthest away from the labour market and contribute to the eradication of poverty”. Such wording sounds perfect in ethical terms.
But what does it mean in practice? The answer is provided by the tool and its real use: the corresponding monitoring indicator, called “New start” is calculated as being the “share of young/adults becoming unemployed in month X, still unemployed in month X+6/12, and not having been offered a new start in the form of training, retraining, work experience, a job or other employability measure” (see EES 2006 Guideline “Ensure inclusive labour markets”) (European Commission, 2006). National implementation, aimed at increasing performance, puts incentives and pressures on the unemployed to take any available task, whatever it is and as soon as possible. The European definition of what to count as a “job” is rather vague and extensive. It leaves room for free interpretation at national level allowing for the inclusion of new schemes. Maximizing such indicators cannot really improve inclusion in labour markets: it mostly increases precariousness.
A Cooperative Game Between Rational Actors (the Member States and the Commission)
It follows that the EES operates as if it was a cooperative game between rational actors. Such a game sounds like this. Its mechanism is familiar to economic theory. Take the Commission and the Member States as the players. The aim of the game is to maximize the key indicators, those intended to evaluate the policies being followed. Actors know in advance the formatting of future evaluation of their actions. Insofar as any learning outcome takes place, it is of a rational order and likely to affect the procedure. Cooperation consists, for each Member State, in manipulating the rules of its own measures and their implementation to meet the requirements of European indicators. In the cooperation, there are invisible but known conventions between actors not to go beyond what each actor was ready to accept. It is not a collective action aimed at genuinely improving employment in Europe. Due to the limited competences given to the European level, Member States are not held responsible for a substantial improvement in European employment, nor do they feel themselves accountable to such improvement when they define their employment policy actions and coordinate with the others in the EES framework. The only constraint is that they have agreed—and this commitment derives from the management by objectives of the OMC—to be accountable vis-à-vis the Commission with regard to their national scores over the whole set of indicators.12
This whole process fabricated positive quantitative outcomes. The global rate of employment (in the European definition) has risen between 1997 and 2005 for the three countries: +2.9 for France; +3.8 for Germany and +2.1 for the UK. Table 12.1 tries to compare these results with the evolution of a full-time equivalent rate of employment. It corrects the global rate of employment with the decrease of the annual number of hours effectively worked by person between 1997 and 2005. The difference between the two roughly estimates the impact of the increase of short-term and precarious jobs, among them the subsidized schemes of return to jobs (for instance the mini jobs in Germany): +0.7 for France; +3.4 for Germany and +2.7 for the UK. Beyond approximations, the impact is notable, more important in the countries already engaged in the move like Germany and the UK, than in France which at that time appeared reluctant. The computation made by a team of researchers (Bruyère et al., 2006, pp. 363–370) was overwhelmingly difficult (in particular for hours effectively worked by a person). To my knowledge, it seems that such an undertaking has not been renewed, though it would be extremely relevant.
Table 12.1
Trends in the overall rate of employment (age 15–64), 1997–2005, in France, Germany and the UK
1997 | 2004 | 2005 | |
|---|---|---|---|
EUROSTAT employment rate (from Community Labour Force Surveys) | |||
France | 59.6 | 63.1 | 63.1 |
Germany | 63.7 | 65.0 | 65.1 |
United Kingdom | 69.9 | 71.6 | 71.7 |
OECD employment rate1 (from national accountings) | |||
France | 60.2 | 63.3 | 63.1 |
Germany | 67.3 | 71.0 | 71.1 |
United Kingdom | 70.2 | 72.3 | 72.3 |
Annual number of hours effectively worked by person2 (from both Community Labour Force Surveys and OECD) | |||
France | 1559 | 1531 | 1542 |
Germany | 1537 | 1468 | 1464 |
United Kingdom | 1697 | 1631 | 1635 |
OECD adjusted rate of employment (corrected from the evolution of hours worked by person from 1997) | |||
France | 60.2 | 62.2 | 62.4 |
Germany | 67.3 | 67.8 | 67.7 |
United Kingdom | 70.2 | 69.5 | 69.6 |
Just a (significant) anecdote to conclude this section: at its own expense, the Belgian employment administration was worried to discover the very low rank of Belgium among European countries in the national benchmarking along the “New Start” indicator. The reason was not the bad functioning of Belgian labour markets, but the Belgian definition of inclusion. To be considered as included in the labour market, the job found must have lasted at least two months. When this was not the case, people remained classified as “unemployed”, which led to a higher registered unemployment duration. The Belgian administration quickly corrected this “mistake” by cancelling this constraint on employment duration. Its quantitative performance improved at the satisfaction of all European and national officers, except Belgian unemployed people who were now compelled to accept any task as a job.13
Quantification: Contrasting Rational Governance with Democratic Choice
Comparing the two phases of the history of unemployment (emergence, deconstruction) offers some incidental views on the differences of democratic choice versus rational governance. In both cases quantification plays a central role, though in very different ways.
Democracy and the Emergence of the Category “Unemployment”
Democratic choice does not consist simply in putting into place optimal procedures for making a choice, or in asking an assembly, even a democratically elected one, to vote. What must be achieved is a free and pluralistic process of public debate, taking the time to weigh all aspects of the choice to be made, without rushing to come to a conclusion. In such a debate, the establishment on the subject in hand of a knowledge basis that should be collectively considered as just and fair is a key dimension, often underestimated. Just in the sense of not forgetting any relevant information, fair in the sense of obeying some shared principle of justice. It is thus, above all, a multifarious social and historical process, driven by many social forces, and not simply the construction of a rational choice operated by the Centre.
Regardless of the country and specific forms that ensued, the “unemployment” category, especially, emerged and developed itself roughly between the 1880s and the 1950s. It was the occasion of a vast and long public debate encompassing contrasting and opposing views, with peak moments at certain points in time. In each country, in its own way, this debate took place in different arenas (political, economic, social, intellectual, statistical) propelled by organizations and their modes of expression (reviews, scholarly societies, public events and demonstrations, etc.). The debate was pursued at different levels and on different scales, in parallel, or in coordinated fashion, within local and regional entities, sectoral, professional and trade groups, and internationally. This process preceded or accompanied the creation of legislation, regulations and institutions. Most of the collective structures where at the time these debates took place were hardly democratic, properly speaking, if “democratic” is taken to mean that the bodies are duly elected and entrusted with a specific mandate to debate issues and propose measures. They were rather the result of a need for collective expression that arose at the time, whether under an authoritarian regime like the German empire or regimes with democratic leanings as in France and the United Kingdom, whether the right of freedom of speech existed or not.
A democratic process of choice cannot be decreed from above, or from outside. This process is often messy, not controlled, nor foreseeable. However, as we said before, democracy is intimately linked with inquiry. The answers people give to an inquiry have an intrinsic democratic value, for they have the capacity to reveal gaps between citizens’ understandings and official intentions for the domain under scrutiny. These gaps underscore disagreements and the plurality of social experience of the same reality, hence the possibility of several relevant questionings, other than the official one. Both between countries and within them, these disagreements emerged as to how to understand unemployment and to count the unemployed.
Due to the plurality of relevant judgements on a given situation, the most important moment in democratic choice processes is not the final step, the decision, but the preceding phase, the reaching of an agreement between actors on the “reality” of the situation, on what is at stake, and on the relevant features to take into account when framing the decision to be made. So, a major democratic concern is to enable people and stakeholders to reach, at least partially, an agreement on the pertinent reality that matters for their choice; that is what Amartya Sen calls the informational basis for judgement in justice (IBJJ), as we will see in the final section further below (Part IV).
Governance-Driven Quantification and “A-Democracy”
The deconstruction process undertaken by European authorities is in contrast with what we might envision as elements of democratic choice in the earlier process of invention of the “unemployment” category. The big change is that, instead of being the fruit of long-term collective debates implying a variety of actors at different levels, the informational bases that pilot the choices are now predetermined from the top by the Centre without any serious deliberation; they incorporate norms into quantification processes before discussion and choice. Such bases orient the decisional processes towards some prefixed types of political outcomes, the ones that the most “naturally” comply with the embedded normativity of the data. These norms are mostly incorporated into technicalities (definition of operational categories; rules of management implementing political schemes; exploitation of the data produced, and so on). Remember the political recourse by the European Commission to ILO statistical categories; and the set of indicators that offer biased models of labour market functioning, or the subtle ambiguity between guidelines and indicators.
Political parties and collective organizations become involved in discussions whose questions, informational bases, and agenda have been prefixed before, on which they have no grip (and often no true understanding of the stakes). Classical representative political democracy and social democracy, too, are circumvented and their role weakened.
I will call “a-democracy” a political regime that maintains the formal procedures of democracy, but impedes, not formal participation of citizens and actors, but any palpable outcomes positive for them (meaning by positive outcomes those that truly improve their situation). Several trends progressively reinforce the efficacy of such a political regime, viewed from the point of view of the political elite and professional politicians. We will point out three aspects of such self-enforcing trends: creating cognitive ambiguity; fabricating quantitative proofs and justifications; generating difficulties to articulate alternative legitimate claims.
Creating Cognitive Ambiguity
The “veil of ignorance”14 surrounding the statistical conventions used to produce the figures creates a situation of cognitive ambiguity. This ambiguity acts like a smokescreen, allowing the conventions adopted as benchmarks for public policy to be changed without any awareness or protest on the part of the public. For example, if the employment rate goes up, ordinary citizens conclude that their chances of finding a job (corresponding to their criteria for a good job) are going to improve. But the European authorities may well—and in fact do—ascribe a different meaning to the notion of employment, one that resonates with the labour market deregulation policy they are pursuing, which obviously works against the expectations of the ordinary citizen. Since it is difficult for citizens, who have nothing but their individual and local experience to test general categories, this situation may last. In a situation of cognitive ambiguity, the task of the authorities consists in maintaining discursive consistency between the established meaning and the new meaning they assign to each category. Public administrations and politicians both are incited to follow this opportunity to maintain such discursive continuity, as it provides them with better justifications. Referring to Austin (1962) (as mobilized by Bohman, 1996, p. 204), one could say that, while employing the same discourse, the European Commission is acting to modify all the possible worlds in which the language convention (“to have a job”) is valid. Believing they have remained in the same world, citizens looking for a job according to the established categories in their world, are confronted by a world in which the same terms are interpreted differently and refer to other actions.
Fabricating Proofs of Effectiveness and Efficiency
What is more, through its self-referential logic, this political method produces justifications of its efficacy that are not only theoretical or discursive but also quantitative. The change in the rules of public policies (employment policies here) does not aim to improve actual social situations but to directly boost scores on performance indicators. The ratings go up without any real improvement in social situations. In fact, those situations may even deteriorate under the impact of standard, short-term measures that cost little per beneficiary because they are designed to affect as many people as possible. The management of public agencies—from the national to the local level—is reorganized according to the logic of performance criteria (Salais, 2010). As a result, the data based on management and on assessing operating rules show progress is being made. They may even be used to demonstrate the veracity of the policy position. In other words, even if it was not their initial goal, reforms tend to establish a direct connection at every level between management and the production of evidence—in other words, self-fulfilling justifications.
Generating Difficulties to Articulate Alternative Legitimate Claims
Creating an environment of procedures of information and of evaluation adequate to predefined political goals (ultimately, a system self-producing proofs) leads to growing difficulties to articulate legitimate alternative claims. As figures and procedures are seen by most of the people as guaranteeing truth by their mere existence, they allow for the endorsing of political credibility. Even if the public debate begins to be fed with such fabricated data (without any professional or democratic control of their process of production), which raise scepticism, it nevertheless means for people that the “facts” are already there. As already existing evidence, these “facts” format the public debate. So it becomes harder to set claims which have not been the object, not only of cognitive elaboration but, more deeply, of common knowledge. For to be heard, claims need to be backed by other socially produced facts; facts that could constitute the basis for shared understanding within the political community and can successfully contest the “official” facts. Following Dewey (1927), such understanding should not be purely intellectual, but also embedded into the engagement of people into “publics”.
If not, the path for democratic expression is cut, even if, formally, democracy remains. The social foundations for active political participation and of citizenship would be undermined, the value of them disappearing for a growing part of the population. By the same process, quantifiers and their demanders are trapped in self-referential loops in which data is taken as the right mirror of reality and, finally, as the reality itself. So the “real” disappears below its quantitative representation which, being taken as the true real, becomes the basis for defining and implementing management reforms and, more generally, public policies. But losing a grip on political and social reality is dangerous for the political credibility and effective performance of policy makers and politicians. A-democracy is the ultimate step of the diffusion of such political methods.
Such political trends call for alternative solutions that correct their negative outcomes. What needs to be put in place to ensure pluralism? What about those who are vulnerable (e.g. citizens with disabilities who cannot easily articulate their opinion)? What should be the relationship between lays and experts in such debates? How far should participation go? How should deliberation be organized? And what are the pitfalls? These questions are beyond the scope of this chapter, mostly because they are waiting for a relevant effective political agenda that does not yet exist. To be possible, it requires, above all, collective learning on the subtleties of social processes of quantification. The first step, in our view, is to be able to develop an approach to quantification that is open to the social critique of its use in governance issues. This is the object of the next section (Part IV).
Social Criticism, Justice and Plurality of Quantification Regimes
The avenue taken by most of the social critics today is the Foucauldian one, especially in English language literature. To quote only one, the work of Wendy Brown (2015) is exemplar. Her book develops a radical and implacable criticism of all aspects of the turn towards a new political governance. At first glance one cannot be but in close agreement with her title “undoing the demos” and her arguments. I discover at work in the governance-driven processes of quantification what I call a-democracy, that is the progressive remoteness of the demos from any effective participation in collective choices. But is it the same as “undoing the demos”? Brown’s subtitle, “Neoliberalism’s Stealth Revolution”, and her demonstration of the omnipresence and omnipotence of neoliberalism leave no room for any collective reaction, or for any counteracting possibilities. Why to exclude any possibility for the demos to survive and find issues? I would like to suggest that our approach to analysis of the relationship between quantification and democracy helps to clarify the point.
In our view, social criticism today must cope with a new element: the emergence of political strategies whose effectiveness lies in acting through the choice of “optimal” informational bases of judgement. Such strategies are perverse, because they distort collective choices in favour of the interests of the central power (and its supporters) at the detriment of citizens and communities. The main worries are that citizens and communities’ aspirations and needs are not correctly represented by the categories, methods of inquiry and data produced that construct the informational bases used to pose and solve collective choice. Above (see Part II), we became aware of such distortions in the case of employment in Europe. European authorities modified the meaning of what should be counted as employment, chose an informational basis centred on the rate of employment and its maximization, all of this pushing the deregulation of labour markets and job precariousness, without any public debate. Evidence is that there is a denial of democracy, biased participation in collective choice and social injustice (with regards to peoples’ aspirations).
Introducing Justice and Democracy
The only way to cut the Gordian knot is to introduce preoccupations with social justice into quantification matters. In every collective choice implying human activities, two objectives should be involved: economic efficiency and social justice, and not only one, efficiency, as in rational governance.15 These objectives should be considered to be at the same level of importance. It follows that, in one way or another, people submitted to quantification in some domain should be asked, or inquired, or adequately represented by movements, associations, political parties at the collective decision levels, on what they consider as social justice for them. This is not a simple thing. We all know—because we experience such moments of feelings of justice or injustice—whether in given circumstances or activities we are well treated or not by others (or by the institution we are facing). But to jump from such personal evaluation to a general principle of justice that would be agreed or accepted by all is another matter. It, no more, no less, requires democracy, an effective one in the making of collective choices. So justice and democracy cannot but go hand in hand in quantification processes. As we have seen above (in Part II), the historical emergence of unemployment statistics in France, Germany and Great Britain reveals some presence of such requirements of justice and democracy that have had unequal collective expressions due to national specificities. Furthermore, in each country several principles of justice competed with each other and had to search for compromise or, at least, for some unstable coexistence at the national level as, in Britain for instance, the deserving poor, the acknowledgement by peers, the morally regular worker.
Remembering such past circumstances today does not mean that the past was better in itself; all the more as social, economic, political realities as well as the people themselves have changed. It nevertheless underlines—not a small thing—that true participation of people, taking into account (to a varied extent) their say and experience of the domain object of public policies, fortunately, is possible. These cannot be excluded. It follows that it is no longer enough today to denounce the governance distortions that are both unjust and non-democratic. One must produce alternative data founded upon just and correct representations of situations and aspirations of people. A different quantification on the same issue should be achieved, based upon another collective “understanding” of the problem to be dealt with. Such quantification has to become legitimate in terms of both fairness and correctness of the data produced; and these data are to be offered to public debate in all their dimensions. Becoming objectively and politically legitimate is the necessary condition to be accepted in the public debate and to be opposed to the “official” basis promoted by the Centre. There is, at the same time, a need to develop a collective social movement able to take charge of the process and to oppose the Centre.
The “Informational Basis of Judgment in Justice” (IBJJ)
The only economist (and social philosopher) that I know for his deep concern about social justice in quantification matters is Amartya Sen. There are others, however, in my view Sen’s works are the most appealing and enlightening ones for us to go further.
The crucial point in Amartya Sen’s approach lies in his emphasis on the informational basis of judgement in justice (IBJJ), which determines the content and methods of collective choice in a democracy. Sen maintains the need for an objective assessment of the state of persons (against the dominant trend of purely ordinal rankings in theories of justice). Sen’s accent on objective assessments connects his approach to quantification issues. Sen introduces in these issues, as soon as human beings are involved, the need to provide as grounds for agreement between people (and for disagreement, as we shall see), tables and indicators that must be just, in the twofold sense of objectively right and socially fair. If so, tables and indicators will cover what, in a genial intuition, Sen calls “the factual territory” over which considerations of justice would directly apply:
The informational basis of judgment identifies the information on which the judgment is directly dependent – and no less important – asserts that the truth or falsehood of any other type of information cannot directly influence the correctness of the judgment. The informational basis of judgment of justice thus determines the factual territory over which considerations of justice would directly apply. (Sen, 1990, p. 111)
This definition of an IBJJ has been introduced by Sen in the context of a dispute with Rawls within the theoretical field of theories of justice. I will just say a brief word on this debate. Sen argues that:
Interpersonal comparisons that must form a crucial part of the informational basis on justice cannot be provided by comparisons of holdings of means to freedom (such as “primary goods”, “resources” or “incomes”). In particular, interpersonal variations in conversion of primary goods into freedom to achieve their life objectives introduces elements of arbitrariness into the Rawlsian accounting of the respective advantage enjoyed by different persons; this can be a source of unjustified inequality and unfairness. (Sen, 1990, p. 112; italics in original)
It is worth noting that for Sen the freedom to achieve should be an actual freedom, not simply a formal one. People should have access to means calibrated to offer them true possibilities, though it is up to them to realize these possibilities, or not. It implies that quantification objectives and methods cannot be but defined in coherence with the objectives and implementation rules of the corresponding policies (see also Salais, 2008).
One will not follow Sen in his debate with Rawls further. Their principles of justice are ones among others. But we will insist on the tight connections with our discussion on quantification. While it was not the direct purpose of Sen, in practice he severely questions the concepts of “fact” and of “objectivity” as usually understood and implied in governance-driven quantification. Most often, the fact is reduced to the status of evidence, something that is not contestable. For a given problem in a given situation, there is only one valuable set of facts, those that pass the test of evidence. No need for justice considerations. By contrast, Sen demonstrates that to be truly objective, an informational basis—in other terms a quantification—should satisfy criteria of fairness (like, in his case, “the actual freedoms enjoyed by different persons–persons with possibly different objectives—to lead different lives that they can have reasons to live” (Sen, 1990, p. 112)). Thereby, in introducing the notion of factual territory, Sen implies that for a given problem in a given situation, there can be several different factual territories, depending on the principles of justice that are applied. It follows, first, that all these factual territories are a priori valuable for posing the terms and purposes of a collective choice on the issue at stake; second, that to evaluate how far the data produced are right requires two things, that they have been produced along rigorous methodologies (correctness) and, too, that judgement and agreement (or at least satisfying compromises) have been achieved between the involved persons and actors on the chosen principles of justice (fairness).
Deliberative Inquiry as Data Processing
The fecund intuition of Sen regarding deliberation from the point of view of social criticism is what is at stake is not prior deliberation over which norm is the right one (a conception based on a hypothetical ontological plurality of norms), but deliberation suited to an adequate grasp of the social reality (a conception based on the observation of a variety of situations from the point of factual territories of justice). Due to the impossibility to objectively decide between ontological norms, an approach in terms of plurality of norms falls into an endless “reconciliation through the establishment of justificatory equivalences” in line with Boltanski and Thévenot (2006 [1991]). In so doing, as Pellizzoni rightly points out, social criticism becomes unable to pose any foundational opposition. Especially, to return to our object of analysis, it would fail to address the “regimes of truth” established by governance-driven quantification. As we have seen above (in Part III), such regimes of truth are precisely fabricated so that “even contesting parties are compelled to accept [them] and to channel their dissent within specific boundaries and on a specific plane” (Pellizoni, 2012, p. 10; see also Pellizoni & Ylönen, 2016; and the conception of deliverative inquiry in Bohman, 2004).
It follows that social criticism should give priority to building social facts that, fairly and correctly, represent the territory of justice that the community judges relevant to the collective objective under consideration. Considering the variety of these territories for the same collective objective, the search of the relevant levels to build these facts, the cognitive categories to be used at these levels and the methodologies of inquiry are open questions to be posed and solved. As the members of the community possess the ultimate practical knowledge of the concrete reality of situations, they themselves only can provide access to what remains inaccessible even to the smartest researcher or observer, the data coming from their experience of the situation. Without their participation, it would be impossible to bring out—or to closely approach—the complete internal and external relevant features of their “factual territory”. These data are not evidence reflecting reality; they are elaborated by people through the prism of their own feelings on what is or is not justice and injustice.
It means that access to such data is not only a question of inquiry in the classical social sciences conception; it has to do with an “extraction” from the people of intimate practical knowledge that they know without knowing that they know it; which means that they should deliberate with researchers all along in the process of inquiry. Such inquiry should be defined as a deliberative inquiry. Its specificities are that its levels, cognitive categories and methodologies, as well as its participants should be “produced” along the processing of the data itself. There is no a priori standard recipe, but something multifaceted (mobilizing people, reflexive awareness, political and scientific) to invent collectively.16
Claiming for Another State
While those developing counter-quantification processes may be not fully aware of their expectations, at the horizon of their action is the perspective of another type of state. Let us return to the two sides of quantification processes discussed in the first section (Part I), the “quantifiers” and the “quantified”. Two correlated questions, political and methodological, have to be addressed: the conception and legitimacy of the authorities who lead the process of quantification (the “quantifiers”); and the nature of the deliberative process that surrounds the quest for answers of quantified people. In the context of a plurality of possible data buildings, what one could call the cognitive moment appears more complex than the simple technical administration of some questionnaire or pure imposition from above. To what extent and how do the quantified have some voice in the choices? How far should the cognitive moment be understood as belonging to a deliberative process? These questions largely remain terra incognita, and they require the possibility of a plurality of types of states. We have already seen in above (see Part II) that the respective role of expectations about state intervention versus collective autonomy differed between France, Germany and the UK for the quantification of unemployment and associated policies. France has the most interventionist top-down state, imposing the same rules to all levels. Germany is historically more open to collective autonomy and diversity at the lower levels (Lander and cities for instance) and the UK is navigating in between. European authorities adopted a French-type interventionist style when they imposed the same panel of indicators to all countries for liberalizing the labour market.
In Salais and Storper (1993, 1997),17 we tried to formalize several types of state supported by different conventions between persons and actors. Such conventions allowed us to understand historical examples. Applying a conventions approach (see also Diaz-Bone, 2018; [2015]; Eymard-Duvernay, 1989; Lewis, 1969) means that these types of states are realized, renewed and made stable through common expectations between people and the authorities. They hold by the virtue of shared beliefs that become deeply rooted in institutions. Such an approach helps to define, at least,18 two types of quantification processes, depending on the state that is object of mutual beliefs. It is worth noting that, if one “partner” (quantified or quantifiers) moves towards another convention of the state, political tensions and conflicts arise. A road is potentially open to social criticism for claiming other public policies, provided it organizes its counter-quantification around another convention of the state than the one already implanted.
In the first convention of the state,19 evaluated people devolve to the central authority the whole task of building the quantification process (modalities, what and how to measure). One can imagine several ways to legitimate such devolution: such tasks are accepted as technical, so no need for voices to be expressed (the European conception again); or, through their representatives, evaluated people are asked to indicate if they agree with the choices made by the central authority. The applied procedure is similar to the one which is used in standard representative democracy. But, for Europe at least, are we still in a democracy or in a move towards what we call a-democracy? Such a convention seems today being replaced by governance by numbers and a-democacy. In such a regime of truth, objectivity is reduced to standardization (Porter, 1992).20 As we have asserted for Europe above (see Part II), in practice evaluated people have no say on choices on the informational basis (the set of indicators); they cannot be truly committed to take the evaluation procedure as their practical benchmark. In a-democracy, such a question becomes irrelevant, because the problem is no more to achieve an effective substantial evaluation, but only to betray current beliefs and representations by producing data apparently supporting them.21 In contrast to the following second convention, there is no need for true deliberation in a-democracy.
In the second convention,22 the authority and evaluated people choose to build a part or the whole of the procedure together, including questions of what and how to measure issues. In practice, it requires that both sides commit themselves to deliberative procedures, which are aimed at achieving deliberate decisions. Such a conception of the state is for us the most fitting for social criticism developing counter-quantification. In contrast to strategic decisions obeying instrumental rationality, deliberate decisions are decisions that both sides have the effective intention to afterwards apply. One will not go further, except to note the proximities with the concepts of subsidiarity23 and of deliberative democracy.24 People should have their say and be mobilized for imposing their views. One cannot expect from central authorities that they spontaneously enter into such a demanding coordination. In his works, John Dewey (1927) has explored the political conditions making such frames of coordination possible more in-depth. Dewey understands democracy as a collective practice led by collective movements that struggle for creating what Dewey calls publics. Publics are to be built along a process that progressively gathers people together to defend a cause (a common good for instance). But such a process is not political in its standard understanding. Political movements mostly conceive such a process as based on ideological or strategic arguments. For Dewey, it consists of a collective learning process anchored in the collective search for the knowledge relevant for implementing the cause at stake. It is, more or less, for people the search for their “true” common world in our pragmatic meaning of the concept. The ultimate stake for them remains not only to publicly oppose their understandings and proposals to those of the authority they are confronted with (which is necessary), but also basically to generate in their community (also necessary) whatever it is, an openness towards conceptions, pragmatic compromises or agreements taking on board their true common world.
Conclusion: Implications for Research on Quantification Processes
The development of governance-driven quantification processes creates opportunities to have a fresh look at factors which previously were taken for granted and not considered problematic. For they introduce to the fields of research and social practice of quantification new concerns about democracy, participation in collective choice, and social justice. The possibility of a plurality of “data makings” for the same situation becomes now visible, thanks to the different relationships of social cognitive practices to politics. Where are their respective scientific and political legitimacies? Should we consider the potentiality and even existence of a plurality of quantification regimes? In line with Sen’s conception of informational bases of judgement, introducing considerations of justice into quantification processes should become relevant and, even more, necessary for better efficiency. One knows how far the right coordinationbetween people depends on their expectation to be fairly treated by others and by institutional or regulatory frameworks that surround their activities. There are several principles of being fairly treated, in other terms of justice. If such an assumption of plurality is relevant, it would extend to the objectivity of data. It also means that a regime of quantification can be validly contested by another one; such contestation should be conceived as a necessary component of any democracy. It opens the road to social criticism based on the creation of alternative informational bases, all being politically and scientifically relevant and legitimate.
Notes
1.
For instance, in his remarkable, internal and procedural analysis of the OMC and its impact on national social policies, Zeitlin (2009) never mentions the impact on quantification and evaluation.
2.
In that respect, the USSR and the People’s Republic of China appeared as pioneers in developing such utopia. See the contributions by Tong Lam and Martine Mespoulet in this volume.
3.
See below the subsequent section which discusses the domain called “unemployment”. What does it mean to be “unemployed”? The official understanding today and everywhere make reference to the ILO definition: actively searching for a job; having no job; to be immediately available to take a job. One will recall that, historically, and depending on the country, to be unemployed was not clear and took time to be so for people.
4.
An example is, in France, the fact that, until the 1950s, female homeworkers, though knowing periods of no work each year, did not produce in the population censuses answers allowing to classify them as “unemployed”. Similarly urban craft workers did not register them as unemployed at manpower bureaus, considering this as an insult to their dignity.
5.
Here we draw lessons from a series of researches, starting independently from each other in the 1980s. See here in particular Phillips and Whiteside (1985), Salais et al. (1986) (reprinted in 1999); Keyssar (1986), Piore (1987), Luciani (1992), Mansfield (1992), Topalov (1994), Mansfield et al. (1994), Whiteside (2007, 2014), Zimmermann (2001), Salais (2011) and Latsis (2006).
6.
As demonstrated by the example of craft workers who do not register in unemployment bureaus, but have their own systems.
7.
See also Storper and Salais (1997) and Salais (2015). One takes this opportunity to rectify a misunderstanding in Thévenot’s contribution to this volume (see note 15) who speaks of “some familiarity with orders of worth”. The foundations for our worlds of production have not much to do with those of orders of worth. They are centred on the product, at the crossing of production and market, precisely two basic economic principles (economies of scope vs economies of scale for the productive organization; risk vs uncertainty for the market; and not on disputes). Furthermore, the state is present as a specific convention with regards to the common good. The only resemblance, is the use of pluralism, which is a brand mark of the economics of convention since its beginning. We already used it in Salais et al. (1986).
8.
9.
10.
11.
Here I disagree with a radical Foucauldian interpretation, which is inclined to see the paw of the monster Neoliberalism everywhere. See also the last section of this chapter (Part IV).
12.
This analysis can be found in Salais (2004) and in Salais (2006). It took time for political scientists specialized in the European domain to understand the complexities of the game. They took the EES as if it were a purely political procedure, with virtual disregard for other factors (especially for the status and formats of numbers). Most of the studies have focused on the wide range of actors for whose involvement the European texts contain provision and on the procedures laid down to organize their complex interactions; this is the famous “multi-level governance”. In the English literature, studies of such gaming and ranking can be found in Bevan and Hood (2006), Hood et al. (2008) and Hood and Dixon (2010).
13.
Raveaud and Salais (2002) analysed all the problems connected to the calculation of European Employment indicators. A more detailed draft is available on request from the author.
14.
To draw on Rawls’ famous concept, which is well suited for the issues described here.
15.
NPM defenders would also say that they are not only concerned with efficiency, but also with effectiveness and outcomes, i.e. to what extent performance meets the stated objectives of a policy, which can include objectives of enhancing equality, fairness, etc. The problem, however, is how such objectives are then made “governable”/measurable through indicators that are quite removed from the original goals (as we have shown before for the example of unemployment in the EU).
16.
17.
18.
19.
This conception corresponds to the convention of the external state.
21.
Michael Power (1997) developed the same conjecture for audits, namely that they mostly support current beliefs.
22.
Which corresponds to the conventions of the situated state.
23.
Acknowledgements
My many thanks for helpful comments by Andrea Mennicken and Ota De Leonardis.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.