Measuring Welfare beyond   GDP – ‘ Objective ’ and ‘ Subjective ’ Indicators in Sweden , 1968 – 2015

This article analyses a series of negotiations on how to measure welfare and quality of life in Sweden beyond economic indicators. It departs from a 2015 Government Official Report that advanced a strong recommendation to measure only ‘objective indicators’ of quality of life, rather than relying on what is referred to as ‘subjective indicators’ such as life satisfaction and happiness. The assertion of strictly ‘objective’ indicators falls back on a sociological perspective developed in the 1970s, which conceived of welfare as being measurable as ‘levels of living’, a framework that came to be called ‘the Scandinavian model of welfare research’. However, in the mid-2000s, objective indicators were challenged scientifically by the emerging field of happiness studies, which also found political advocates in Sweden who argued that subjective indicators should become an integral part of measuring welfare. This tension between ‘subjective’ and ‘objective’ measurements resulted in a controversy between several actors about what should count as a valuable measurement of welfare. As a consequence, we argue that the creation of such value meters is closely intertwined with how welfare is defined, and by what measures welfare should be carried through.

(GDP).This complexity, and the lack of agreement upon what counts as welfare and what does not, has resulted in an absence of standardised units or scales that could be used for cross-country comparisons.Even though there have been many proposals of indicators for measuring welfare, the measurement of welfare beyond GDP has often been contested at the national level, and has for the past half-century been the subject of major revisions in most modern states.
In the 1970s, a new approach to measuring welfare was pioneered by the Social Indicators Movement.A new generation of social scientists began questioning GDP and Gross National Product (GNP) as valid measures of welfare, at least when used as exclusive indicators (White 1983;Cobb and Rixford 1998).Instead, the Indicators Movement started defining various self-reported measures of 'subjective' well-being and quality of life, which actually relied on 1 scales that could easily be administered in surveys, such as Hadley Cantril's Self-anchoring striving scale (Cantril 1965) and Norman Bradburn's Affect-balance scale (Bradburn 1969).Even though the notion of subjective measurements as an indicator of a good society and government dates back to at least Jeremy Bentham's late eighteenth-century idea of a 'felicific calculus' (Bentham 1823(Bentham [1789]]) -an enlightenment vision of a scientific measurement of pleasure and pain that would guide the legitimate governance of a societysubjective measurements were introduced by the Social Indicators Movement in a slightly new key in the 1970s.
In Richard Easterlin's now famous 1974 article (Easterlin 1974), today commonly referred to as the 'Easterlin paradox', and in the works published in the Social Indicators Research journal (launched the same year), GNP as a measurement of welfare was thoroughly criticised.The Easterlin paradox would become a recurring question in the emerging field of happiness studies, and it sparked a scholarly debate on the relationship between economic growth and subjective well-being.Easterlin argued that empirical data showed that economic growth only increased happiness up to a certain level.People living in rich countries were in fact happier compared to people living in poor countries; however, within a single rich country, further economic growth (measured as GNP) did not seem to cause increased happiness.Instead, Easterlin concluded that 'higher income was not systematically accompanied by greater happiness' (Easterlin 1974: 118).
Easterlin's paradox is not the only critique of purely economic measures of welfare at this time in history.One of Easterlin's contemporaries, Tibor Scitovsky, published the widely disseminated book The Joyless Economy in 1976 (Scitovsky 1976), which critiqued the view of a rational consumer in neoclassical economics and showed how American society created hard-working, increasingly consuming, but quite unhappy citizens that did not experience any real improvements in the general standards of welfare.
However, in the 1960s the crisis in welfare economics had already begun, as manifested by the 1961 United Nations' interim guide to define 'levels of living', as a set of indicators for assessing aspects of welfare beyond economical measurement, for example, education, health, nutrition, freedoms and housing (United Nations 1961).This indicator approach broke abruptly with the welfare economics tradition since it downplayed economic measurements as merely being one of several.Instead, it brought forth a set of 'objective' indicators of actual living conditions, combined with indicators that were more difficult to compare between nations, for example, 'human freedoms', which were regarded as varying between cultures.
The indicator approach and the quest for alternative measures of welfare had gained momentum in many countries also outside the United States.Noll and Zapf (1994) summarise the changes in social reporting in post-war Europe as: Whereas the Scandinavian countries, Great Britain, the Netherlands, France and the German Federal Republic were among the trendsetters in the establishment and institutionalization official reporting, the Southern European nations were latecomers.The "classics" among social reports-the British "Social Trends", the Dutch "Social and Cultural Report" and the French "Donnés Sociales"-have now [1994] been published regularly for more than two decades.(Noll and Zapf 1994: 5) The search for the best possible measure of welfare continues to date.Besides economics, a wide range of social sciences has inquired into what welfare is composed of, how it should be measured and what values to bring into it.At the turn of the millennium, the fields of Social Indicators Research along with 'happiness studies' had presented increasing amounts of research on subjective well-being and life satisfaction, as subjective alternatives to the established practice of measuring welfare as consisting of predominantly objective indicators.From the fields of psychology and sociology came proposals for socalled subjective factors to be included as national indexes (see for example, Diener 2000), which in turn relied on a line of research that had emerged over the past four decades (Kullenberg and Nelhans 2015).As we will show, this 'subjective turn' would impact policy in Sweden in 2015, as new measurements of welfare were investigated in a government report commissioned by the Ministry of Finance, which serves as the locus of this article and as entry point into the controversy between subjective and objective indicators of welfare.However, to analyse the role of such measuring devices of welfare, we will consider a few theoretical implications in the study of what is sometimes referred to as 'value meters'.
Theoretical Perspective: The S tudy of ' Value Meters' In this article we wish to draw special attention to certain instruments or devices in modern societies that both render society and the economy knowable, simultaneously with their being acted upon in political decisions.We will analyse the negotiations of establishing a measurement of welfare as being performative-in the sense that the act of measurement co-creates the social order that it was designated to describe.In early actor-network theory (ANT), the institutions that performed such measurements were often referred to as 'centers of calculation' (Latour 1987(Latour , 1999; see also Czarniawska 2004) that had the double function of producing statistics used for counting and acting upon the social world (Saetnan et al. 2011).More recently, the notion of value meters has been advanced by Latour and Lépinay (Latour and Lépinay 2009, Latour 2013, also called 'valorimeters' in Latour and Callon 1997) to designate devices of calculation that have a different functionality compared to scientific instruments.Whereas 2 scientific instruments create chains of reference in order to reach remote states of affairs, value meters instead designate devices that 3 both measure something and take action to change what they measure.They are especially prolific in economic matters, and are often the products of economic research: A similar concept is 'market devices', which consist of 'the material and discursive 2 assemblages that intervene in the construction of markets' (Muniesa et al. 2007: 2).Here Muniesa et al. show how market devices have the dual function of describing and acting upon the market.However, and this is more a matter of detail than of theoretical significance, the notion of value meters is somewhat more flexible for analysing non-monetary transactions that do not immediately act upon the market (such as the measurement of subjective feelings).
This idea has remained more or less intact throughout much of Latour's writing.
3 However, its most recent formulation in An Inquiry into Modes of Existence (Latour 2013, esp. p. 77ff) integrates the scientific chains of reference into a philosophical system.This way, the contrast with other modes of veridiction becomes much clearer (for example those of legal systems, political speech or technology), and this is also why the notion of 'value meters' in this work can be contrasted more efficiently compared to previous work, and can be compared to other 'valuation devices' more easily (see Martinus Hauge 2016: 128-9).
When the disciplines of economization arose, they manifested themselves by the overabundance of these quite particular types of "quali-quanta" that connect the two senses of the French expression prendre des mesures: "taking measurements" and "taking measures"-hence the term "value meter."(Latour 2013: 408, emphasis in original).
Value meters, such as stock quotes, balance sheets, scholastic performance in school pupils (for example the PISA scores), indexes of economic growth (for example GDP) or fluctuations in the price of natural gas are measured because they inform a decision about whether to take a certain measure or not.Stocks are bought based upon their price; budgets are based on balance sheets; pedagogics on the latest PISA results; and the decision to begin exploiting energy resources is based upon the current price of natural gas.The measurements do not necessarily need to be monetary, but rather, value meters are metrological in character and when 'connected together, little by little, end up building metrological chains which make the inter-comparison of subjectivities increasingly "precise," "accentuated," and "worthy of being objects of speculations of a new sort"' (Latour and Lépinay 2009: 19).
Value meters thus imply both the quality of what is measured-its value-and the quantity of that value, which in turn makes calculation possible.They invent, so to speak, their own measure of what should be measured.In this article we will follow the construction of value meters that measure welfare.We will analyse them prior to their becoming stabilised, in their defining moments, as they are still a matter of debate, just before they turn into common usage and become veiled under a curtain of perceived objectivity, almost as if their qualitative definition of value and their mode of quantifying that value had always been there.Precisely because value meters are so integrated into 'society' and the 'economy', the design choices that were made in order to construct them seem to disappear in their everyday usage.
In this article we will attempt to analyse value meters from two sides.As Zuiderent-Jerak and van Egmond (2015) argue, market devices (or value meters) should not be reduced to what explains a certain order.In addition, they also have to be explained from the (cultural) context from which they emerge.As our study spans several decades, we will try to alternate between providing an account of both the historical context that make certain calculations possible and how these devices, once they are constructed, render new values quantifiable.

Pur pose and Researc h Questions
The purpose of this article is to analyse the disagreement on how to measure welfare in Sweden.We will draw special attention to a controversy between what sociologists call 'subjective' and 'objective' indicators in social reporting.This way we will analyse how the measurement of quality of life and welfare beyond the GNP/GDP can be stabilised along such lines, depending on the outcome of negotiations taking place in science and politics.We will focus on the 'Scandinavian model of welfare research', particularly the Swedish variety, and trace how it has been both challenged and strengthened from its early formulations in the late 1960s up to today.The goal, however, is not to present a complete history of social indicators in Sweden.Instead, we will approach three defining episodes of negotiations about what counts and what does not count as a measurable value in relation to welfare; in short, what we discussed earlier in terms of value meters.In other words, we want to understand how welfare has been valued over and above economic measures and calculations, and how such values have been perceived as key components in social progress.We will proceed by identifying a controversy concerning the measurement of welfare, a question that has been expressed both in political terms as well as through social scientific reasoning, and has appeared several times during a time span of almost half a century.We have selected three episodes and sites of controversy, where the first two instances converge in the third trial in 2015, which has not yet been settled and remains an open question to date.
We will enter the debate on indicators of welfare in the late 1960s to give a background of how Swedish sociologists were commissioned by the state to conduct surveys and write reports on 'levels of living'.In their inquiries they defined the 'Scandinavian approach' to welfare indicators, which relied exclusively on what they called 'objective indicators'.Then we will move forward to the mid-2000s and analyse a series of political proposals made by the Swedish Green Party (Miljöpartiet) in which it was suggested that the government ought to measure subjective well-being and happiness instead of GDP as indicators of welfare.Here, we analyse how political discourse utilised scientific knowledge as a key argument to envisage a new way of measuring welfare with subjective indicators, in contrast to the previous approach based on objective indicators.Last, we will focus on a 2015 government report, which commissioned one of the early sociologists who contributed to the Scandinavian model of welfare research, Robert Erikson, to reform the indicators of welfare to be used as official statistics by the state.As we will show, this report supports the objective measurement of welfare.The report can be contrasted with international trends where subjective measures of wellbeing and quality of life are today used ubiquitously throughout multiple scientific disciplines, some even arguing that there has been a 'happiness turn' in the social sciences (Ahmed 2010;Kullenberg and Nelhans 2015).Consequently, the 2015 report was met with strong opposition as it was circulated in a round of expert referrals.We will conclude by analysing these discussions, with special attention to how the valuation of the 'welfare of nations' is related to the epistemic practices of the social sciences.

Met hodology and Mater ials
To understand how social scientific knowledge contributes to the construction of value meters in a welfare state, we have collected different types of documents: scientific literature, government reports, think tank reports, proposals, protocols and debates from the Swedish Riksdag, interviews and opinion pieces published in news media and expert referrals made by a wide variety of institutions.As these documents have different ways of telling the truth about their subject matter, what Latour (2013) calls different modes of veridiction, they have to be sorted and analysed with special attention to what resources they mobilise when the respective subject matters become controversial.
To study value meters, we have utilised a network approach to detect specific 'crossings' that have made possible productive entanglements between scientific results and values.We have proceeded according to the well-known principles laid out in Latour's Science in Action (Latour 1987) but with the important addition of extending the definition of a 'document' to also include accounts that originate outside scientific networks (Latour 2013;Verran 2014).Thus, we have expanded the possible links and citations (what constitutes a network) to also include political speech, accounts circulating in the public sphere, reports and institutional records.This way, we are no longer limited to scientific literature as defined in Science in Action, but instead we are able to detect various crossovers between scientific forms of veridiction and other types of statements.
We have utilised a number of databases to collect documents relevant to understanding the process of establishing an alternative measurement of welfare.These include: • the Swedish Riksdag (parliament) open data platform www.data.riksdagen.se • the Mediearkivet (Swedish media archive), a paid service provided by the company Retrieve; • Swepub (Swedish academic publications database), provided by the National Library of Swede; • government white papers (Statens Offentliga Utredningar), which were recently digitised and made available for full-text search by Linköping University (see www.ep.liu.se/databases/sou/); • There were 66 referrals to the 2015 Government white paper (Erikson and Blanck 2015) for creating a new welfare index.The referrals were requested as public documents from the Ministry of Finance (Finansdepartementet) and were delivered electronically.
As a general approach, we used the 2015 white paper (Erikson and Blanck 2015) as the entry point in our analysis.Then we proceeded by collecting as many cited and citing documents as possible (referrals, scientific articles, newspaper articles) in order to generate relevant search strings for the above-mentioned databases.All documents were added to a common database to enable full-text detailed searches.This way we were able to bind together the three empirical sections with citing and cited documents.As a consequence the three episodes, even if they are dated chronologically, achieve their analytical meaning through the connections that hold them together.First, we will go back to the Scandinavian model of welfare research in the late 1960s up until the late 1980s to show how the objective indicators approach was once assembled and made to function in the Swedish welfare state.This first episode serves as the backdrop necessary for understanding the main account that is played out after the turn of the millennium.As a second episode, we will outline a counter-thread of political proposals arguing in favour of subjective indicators made by the Green Party in the mid-2000s.Third and last, we will return to the reception of Erikson and Blanck's report and the opposing sides in the controversy between subjective and objective measures of welfare to arrive at the main controversy between subjective and objective indicators of welfare.

Episode 1: The Scandinavian Model of Welf are Researc h: 1968-87
To understand the controversy between subjective and objective measurements of welfare, and why the former has had difficulties in gaining traction in Sweden (at least as official statistics of welfare), a key thread of social scientific history can be followed back to the late 1960s.At this time, there were no systematic surveys to fall back on, and the Social Indicators Movement had begun expressing fresh new ideas about alternative measures of welfare in an international context.In 1961, the United Nations had published a set of indicators (United Nations 1961), and around Europe, social scientists were responding with various systems of social indicators, which to varying degrees were incorporated into official statistics (Noll and Zapf 1994).The United Nations report discussed 'aggregate consumptions and savings' and similar measurements where national income and consumption was divided per capita to form aggregate indicators.The report argued that such measurements were 'ambiguous' in relation to levels of living and should only be accounted for as 'basic information', not as a proper indicator (United Nations 1961:15-16).
In Sweden, the story of indicators began in 1965 when the Interior Ministry (Inrikesdepartementet) appointed Rudolf Meidner to lead the Low Income Committee, which produced a series of reports with the purpose of increasing knowledge about social groups with low wages.Except for Meidner, who worked at Stockholm University, the rest of the committee consisted of experts from unions (Landsorganisationen and Tjänstemännens centralorganisation) and the Swedish Employers Association (Meidner 1970;Johansson 1973).As part of the report series produced by this committee, Sten Johansson, at the newly founded Swedish Institute for Social Research, Stockholm University, conducted the first Level of Living Survey (Levnadsnivåundersökningen) in 1968, published two years later (Johansson 1970).The survey made use of Richard Titmuss's notion of 'command over resources' as the definition of levels of living, which meant that the individual was regarded as an acting social being.In this account, the individual was seen as commanding a set of resources, such as money, knowledge, physical and mental energy, social relations, etc., which gave him or her the capability to control his or her living conditions, and thus increase his or her freedom in guiding his or her life (Johansson 1970).
Johansson listed several drawbacks when it came to using monetary measurements (such as GDP) in accounting for the welfare of a nation.He argued that such measurements failed to describe the individual's circumstances, as they did not measure the performance of for example students, children or house wives, as these categories did not produce significant monetary gains.Moreover, monetary indicators lacked precision in accounting for local circumstances.The price of the same house or apartment could be radically different if located in a large city or in a rural area, Johansson argued.Similarly, monetary measurements would also misinterpret real-life contexts.Johansson gave the example of commuting in a crowded subway compared to a short walk to the workplace.Only the former made an imprint in terms of GDP, even though it probably was less beneficial to the individual in comparison with a pleasant stroll.Summing up, Johansson argued that monetary measurements were inherently unsuitable for analysing inequalities in society, something which was indeed incompatible with the Low Income Committee's goal of improving levels of living for the poorest citizens (Johansson 1970:17-19).
Three years later, Johansson published an article in Acta Sociologica describing the survey to his sociological peers.He suggested that Level of Living surveys could be seen as part of 'the (now) international "social indicator movement", if that is understood as a push for improved and more relevant social statistics' (Johansson 1973: 212).However, unlike his international peers in the Social Indicators Movement, Johansson opposed what he called 'subjective' measurements and argued that: Concentration on resources, rather than on fulfilment of need, furthermore made questions of individual subjective satisfaction recede in the background.This is of rather fundamental importance, since that aspect determines the nature of the data in relation to the political process.Subjective satisfaction data would function as continuous pseudo-plebiscites in themselves while subjective perception data on resources (not to speak of objective data) can only function as a basis for interest articulation and aggregation for political parties, trade unions and other interest organisations.Subjective satisfaction data would simulate the function of interest articulation of the democratic process in a way that is rather controversial.(Johansson 1973: 213) A returning criticism against subjective indicators, as expressed by Johansson, was the possibility to manipulate the political process.He argued that 'objective' measurements of health, schooling and nutrition could be valuable indicators for politicians.In contrast, 'subjective' measurements of self-reported life satisfaction were 'controversial' because they interfered with the democratic process, as they only measured the (subjective) fulfilment of a 'need'.
The fusing together of scientific measurements and politically relevant indicators was highlighted clearly in Finnish sociologist Hannu Uusitalo's review of Johansson's work: One of the central criteria for the structuring of the components of the level of living has been 'manipulability', that is, the study should concentrate on those problems which can be removed by political decision making.(Uusitalo 1973: 226) The notion of 'manipulability' in this context is perhaps the most crucial difference between scientifically and politically motivated measurements.In the Scandinavian welfare research tradition of the 1970s, 'manipulability' was understood as a criterion of inclusion, with regard to what social problems could be managed by the state and what problems were considered to be out of reach of its welfare system.But, as already mentioned, manipulability with regard to the satisfaction of subjective quality of life, was off limits in a democratic society, according to Johansson and his sociological peers.Manipulating health, housing, schooling or nutrition-the objective indicators of quality of life-should be the task of the welfare state; but manipulating social conditions as a means of increasing the subjectively experienced quality of life of citizens should not be part of the politician's toolbox.The notion of manipulability in this context also shows the duality in value meters of this kind, as opposed to strictly scientific measurements.The indicators of the Level of Living survey were meant to simultaneously measure and transform welfare in Swedish society.
The Scandinavian model of welfare research was, however, not uncontroversial.Johansson's Level of Living survey immediately created political controversy because the results showed that there existed a surprisingly severe level of poverty among the lower working classes.For the ruling Social Democratic Party, this became a pressing issue.Sociologist Casten von Otter, also at the Swedish Institute for Social Research in Stockholm, wrote: It is not often that one gets the feeling that sociology really matters.But for once, with the 1968 Survey of Level of Living in Sweden, a mark is established in Swedish sociology whose effects can be seen also in the political history of the country.Not only has it aroused a controversy of considerable dimension but it has also gone a long way in achieving one of its major aims; to provide the discussion on welfare-politics with a more solid empirical basis.At the same time it relieved us, at least temporarily, of that threat against a welfare-state which is inherent in the belief that we already have achieved it.(von Otter 1973: 229) Johansson's results, together with the Low Income Committee's other reports, had become a hot potato in the Social Democratic Party and Swedish political debate.So hot that the committee was disbanded by the prime minister and the issue became a matter of national concern.Afterwards, Johansson remarked in a footnote that '[t]he political turmoil that followed the dissolution of the Low Income Committee in 1971 has forced some thinking on this problem and on the role of the expert' (Johansson 1973: 213).In retrospect, that was rather an underestimate of the graveness of the debacle.Economic historian Jenny Andersson shows that: [d]uring a historic radio debate, the Minister of Finance Gunnar Sträng greeted Per Holmberg [head of the Low Income Committee] with this question: "What person, sound of body and mind, would work for less than five kronor an hour" The Implication was clear: in a social democratic society of full employment and solitary wage bargaining, only those in some way handicapped could possibly remain poor.In 1971, the Committee on Low Income was disbanded by the Prime Minister and party leader Olof Palme.(Andersson 2006: 54) The Swedish sociologists, Johansson in particular, had not only stirred up a politically sensitive issue of poverty, but they had done so by defining and applying social indicators that went beyond conventional welfare economics, and instead adhered to the emerging social indicators approach (see Figure 1).
The impact of the Low Income Committee continued to reverberate through the 1980s and even received international attention.For example, in 1983, the German newspaper Der Spiegel published several articles about the failure of the Swedish welfare system, and political scientist Olof Ruin argued that the 'Low Wage [income] Committee' was partially to blame for this crisis in the image of Swedish levels of living (Tidningarnas Telegrambyrå 1983).But the political sensitivity of the new value meter seemed only to make it more relevant as researchers attempted to export it as a wider model of welfare research.

The 1980s: Consolidation of t he Scandinavian model
The Swedish Level of Living survey influenced Norwegian, Danish and Finnish researchers who conducted similar inquiries throughout the 1970s (Erikson and Uusitalo 1987;Erikson and Blanck 2015: 76), creating what would be labelled the 'Scandinavian model of welfare research' in international publications.These accounts drew a line of demarcation against subjective indicators into the 1980s, and Robert Erikson, along with his Norwegian, Danish and Finnish colleagues presented this line of research as a coherent entity comparable to other international strands of the Indicator Movement, with only minor differences between the respective countries.A common denominator in the Scandinavian approach was the rejection of subjective quality of life indicators, which were identified as primarily being connected to the American branches of the Social Indicators Movement, in 4 particular the works by Campbell and Andrews, which were described as 'particularly strong in the Unites States … ' [but] ' … [i]n Scandinavia, this variant has fewer adherents' (Erikson and Uusitalo 1987: 185).However, the reasons for relying solely on objective indicators were quite elaborate.
Erikson and Uusitalo argued that Scandinavian welfare research was not primarily driven by academic questions, but instead 'operates on the same levels as policies do' (Erikson and Uusitalo 1987: 189).First, policy centred indicators required '[an] emphasis on applicability [that] is one reason why these studies have used objective and concrete indicators of welfare, since these are the objects of concrete politics' (Erikson and Uusitalo 1987: 192).Second, Erikson and Uusitalo argued that 'objective' indicators were much better suited to study 'inequalities' because marginalised people could also report high levels of subjective well-being, while remaining objectively worse off in terms of lower levels of living.Third, they put forth the argument that subjective indicators represented 'final values', which were particularly unsuitable for a democratic society.If policy were to be guided by, for example, happiness or subjective well-being, it would lead to the risk of a policy that not only allowed for inequalities, but also could be shaped to manipulate state interventions to keep people happy, without improving their lives.Once again, the performativity of the value meter is considered with delicate care, excluding what is understood as subjective elements.
The same year, Erikson and Åberg presented the Scandinavian model to a British readership in the volume Welfare in Transition.
Post-war Swedish sociology, particularly up until the 1960s, was predominantly 4 influenced by American empirical social science, especially in its methodological use of statistics and large-scale surveys, as social scientists were often sent to the United States for their academic training (see Fridjonsdottir 1991;Eyerman and Jamison 1992).
Here, they once again stressed that objective indicators (here called 'descriptive indicators') were preferable in comparison to subjective or 'evaluative' indicators, since the latter fluctuated with the individual's valuation of their own situation and could thus not be said to describe people's actual circumstances (Erikson and Åberg 1987: 4).This problem, often referred to as hedonic adaptation, was a point of departure similar to Easterlin's famous paradox (Easterlin 1974), and is a recurring problem in the field of happiness studies.
For two decades, sociologists at the Swedish Institute for Social Research at Stockholm University had established an internationally renowned '[...] concept which bases welfare measurement exclusively on objective indicators as the level of living approach of Scandinavian welfare research' (Noll and Zapf 1994: 3).This way, a programme for measuring welfare beyond GDP had been created, a programme that adhered strictly to 'objective indicators', instead of including 'subjective factors', which was often the case internationally.This way of measuring society, which fused facts with values and drew a boundary between what could be measured and acted upon and what ought to be left alone to the individual to decide upon for him/herself.Once built into a value meter that in turn could be incorporated into the Level of Living surveys, the demarcation between 'objective' and 'subjective' indicators did not have to be reprised for every new survey, but could persist as a standard measurement device of welfare.this line of research had neither found its way to Swedish politicians nor gained substantial traction in academic research. 6 However, a few years into the 2000s, the critique of GDP as a measurement of welfare returned.But in contrast to the sporadic proposals of the 1970s, this time 'subjective welfare' came to play a central role, as opposed to the 'objective' indicators that had ruled the official statistics for decades.In October 2007, Green Party MP Max Andersson proposed for the first time to the Swedish Parliament that 'happiness research' should be supported by the state, and that a national index of happiness should be created as an additional measure that could be used alongside GDP.Andersson argued: What is interesting about happiness research is the possibility of rendering into numbers various aspects of human happiness.This radically simplifies the task of giving priority and to balance the utility of [political] proposals.(Andersson 2007, our translation) A few weeks later, Andersson participated in a happiness research conference at the University of Gothenburg (Johannisson 2007) and shared his results in an academic setting.Also attending the conference was Bengt Brülde, a Gothenburg-based philosopher who had introduced happiness research to Sweden in books and articles.Andersson's proposal was inspired by recent developments in the UK, where happiness research had been successfully put on the political agenda.Besides Jeremy Bentham, Andersson also cited the influential 2005 book Happiness: Lessons from a New Science by Richard Layard (Layard 2005) and the British Government report Life Satisfaction: The State of Knowledge and Implications for Government (Donovan et al. 2002).When interviewed by the newspaper Göteborgs-posten, Andersson stated that happiness research supported green politics and that a happiness index would be 'uncontroversial and widely accepted in ten years' (Bjärsdal 2015).These accounts, along with Andersson's proposal, all depart from the Easterlin paradox mentioned in the introduction of this article, that is, increased GDP leads to increased happiness, but only up to a certain level of welfare.Instead, most developed countries in the post-war period have ground to a halt where happiness levels remained unchanged, despite a steady growth of GDP (Easterlin 1974).While this relationship is often expressed in policy documents, the scientific community of happiness researchers is still disputing whether the paradox remains valid in light of new empirical studies (see Hagerty and Veenhoven 2003).At the core of the body of research cited by Andersson in his proposal lay the notion of 'subjective well-being', an approach advanced especially by Erikson and Uusitalo mention Finnish sociologist Erik Allardt as a notable 6 exception; he measured subjective components of life satisfaction in combination with objective measurements (Erikson and Uusitalo 1987: 185).psychologists during the 1980s (see Diener et al. 1999).As a measurement scale, subjective well-being surveys had been tested using different scales and survey methods, and had become a widely used form of measurement in psychology and the social sciences.
By following the Green Party proposals, it is possible to detect both a consistency in adhering to subjective measures of welfare and a number of important additions to the original proposal.In 2008, Andersson submitted the same proposal to parliament with two additions.He cited Brülde's recently published book about happiness (Brülde 2007), arguing that Swedish happiness research had advanced to the level of producing textbooks.But even more important, Andersson had kept up to date with a certain commission appointed by Nicholas Sarkozy in France, which one year later would publish the Report by the Commission on the Measurement of Economic Performance and Social Progress, usually referred to simply as the Stiglitz report.Andersson even cited the liberal author Johan Norberg, who had written a report for the free-market think tank Timbro, where he strongly criticised 'labour economist' Layard for a 'paternalist' view on happiness in society (Norberg 2006).Despite being a political foe, Andersson utilised Norberg's report in his proposal as evidence of a 'growing body of literature' on happiness (Andersson 2008).Hence, the political rhetoric of Andersson contains active modifiers, with the aim of assembling even contradictory statements as part of an argument that happiness research has grown into a mature scientific discipline.
In 2009, as the Stiglitz report had been published, Andersson incorporated its critique of GDP as a measure of welfare, and the same year he co-organised another conference at Malmö University, thus adding the number of connections between his political proposal and academic research by discussing the report commissioned by Sarkozy with social scientists (Lindström and Andersson 2009).In Andersson's 2009 proposal to parliament, a specific section of the Stiglitz report was highlighted: […] recommendation 10 suggesting a measure of subjective well-being, often translated less formally as happiness, providing crucial information on people's quality of life, which in turn should be included in official statistics.(Andersson 2009, our translation) Going back to the Stiglitz report, this formulation was originally stated as: Recommendation 10: Measures of both objective and subjective well-being provide key information about people's quality of life.Statistical offices should incorporate questions to capture people's life evaluations, hedonic experiences and priorities in their own survey.(Stiglitz et al. 2009: 16) Andersson excluded the notion of objective well-being in the Stiglitz report and instead narrowed down his proposal only to the measurement of subjective happiness.This modulation towards the subjective aspects of welfare would continue further down the road for the Green Party politicians.In 2010, Andersson left parliament but handed over the task of suggesting alternative measures of welfare to MP Lise Nordin.Nordin did not make any substantial changes to the 2010 proposal, besides shortening it slightly and dropping the reference to Norberg and the outdated 2002 Life Satisfaction report (Nordin 2010).Furthermore, in the same year, Green Party MP Valter Mutt argued in parliamentary debates that Sweden should account for subjective happiness in its official statistics, just as the conservative government in the UK had done, not only to give priority to people's subjective well-being over GDP, but also out of environmental concerns (Mutt 2010).
The Green Party's critique of GDP was in one way not so different from the earlier sociologists' approach.Max Andersson argued similarly that such measurements did not accurately describe the individual's circumstances (Andersson 2008) and that it did not measure welfare, only 'production' (Andersson 2009).However, the point where a major difference appears is when it comes to human happiness.Andersson mainly argued that GDP merely measures the way we are getting richer, not getting happier.Moreover, he wanted to make 'happiness research' on a par with 'economics' in valuating welfare in Sweden (Andersson 2008).
In 2011, Nordin's proposal would both bring in more allies and make new enemies.Previously, UK Prime Minister Gordon Brown (Labour) released a cross-government programme called New Horizons-A Shared Vision for Mental Health (Cross-government strategy: Mental Health Division 2009).The programme was meant to combat mental illness, but it also stressed the importance of wellbeing, which Nordin conceived of as happiness, even though such specific terms were absent from the proposal (Nordin 2011).However, there was another important ally in Nordin's plan.In 2011, the UN General Assembly adopted a resolution saying that 'the pursuit of happiness is a fundamental human goal' and that GDP 'was not designed to and does not adequately reflect the happiness and wellbeing of people in a country' (United Nations News 2011).If the scientific aspect of this new value meter was the first line of rhetoric, the second aspect would be to connect it to certain values, such as human goals, and further down the road, with environmental sustainability.
By now, the Green Party proposals had started to circulate in the public debate.The editorial column of the Swedish newspaper Sydsvenskan reacted to Nordin's proposal the day before it was presented to the Riksdag, arguing that happiness as a political goal, as laid out by both Nordin and Sarkozy, was a 'deeply anti-liberal and paternalistic idea, which should be rejected' (Sydsvenskan 2011).The Sydsvenskan editorial also cited a statement made by Cameron in 2006, saying '[w]ell-being can't be measured by money or traded in markets' (BBC 2006).In an article in the magazine Arena, Kjell Vowles engaged in a thorough round of interviews to cover the current debates on happiness as a political goal.Here, Nordin elaborated on how politics based on happiness research would lead to 'green politics', emphasising sustainable development rather than economic growth (Vowles 2012).This way, subjective indicators on happiness and life satisfaction were connected to political goals of environmental sustainability.This way, the Green Party established the idea of a politics based on 'subjective' indicators, which would in turn lead to fewer concerns with economic growth, thus enabling a greener politics.Even though it was not as thoroughly tested as in the earlier Level of Living surveys, the proposal for a new value meter of welfare had all the principal components more or less ready.The Green Party politicians referred both to scientific advances in 'subjective well-being' research and to heavyweight institutional bodies such as the UN or the UK Government, thus presenting a new composite of facts and values.
However, the Green Party proposals would not be able to gain the support of a parliamentary majority.In 2014, the Ministry of Finance led by the opposing right-wing government, as mentioned above, instead commissioned Robert Erikson, a fierce advocate of 'objective' indicators as we saw before, to conduct the Swedish Government Official Report on measurements of quality of life.

Episode 3: The Year 2015 -The Subjective Measurement Controversy
Almost half a century after Sten Johansson had conducted the first Level of Living survey to measure welfare in Sweden, the Ministry of Finance decided to return to the question of indicators beyond GDP.Even though sociologists at the Swedish Institute for Social Research had continued to work according to their established methods, GDP still held a firm grip on the way the welfare of nations was measured and compared with other nations.Economic growth figures seemed to be difficult to complement with alternative indicators, even if the researchers argued that they were 'objective' in character.
In 2014, the Ministry of Finance, led by the conservative Moderate Party's (Moderaterna) Anders Borg, announced that they were to commission a state committee to create new measurements of quality of life (Finansdepartementet 2014).As head of the committee, sociologist Robert Erikson at the Swedish Institute for Social Research (SOFI) at Stockholm University was appointed (Frisk 2014).The committee was to work quite swiftly.Already one year later, the committee had reported back to the Ministry of Finance.The result, authored by Erikson and his secretary Anton Blanck, was the Swedish Government Official Report SOU 2015: 56 (Erikson and Blanck 2015).
Similarly to back in the days of the establishment of Scandinavian welfare research, Erikson and Blanck critiqued GDP as a measurement of welfare for several reasons.According to the report, GDP did not measure activities that were immediately beneficial for society, nor did it account for social, environmental or economic sustainability in a nation.Moreover, GDP measurement did not detect unpaid labour, for example household work or free information on the internet.In short, Erikson and Blanck concluded that GDP did not measure many things that 'we value in society and the environment we live in' (Erikson and Blanck 2015:27-28).
The Ministry of Finance had specified that the commission should investigate a wide range of indicators, among them also 'subjective' ones, and report back with regard not only to their scientific merits, but also in what sense they were useful as indicators that could inform political decision making (Finansdepartementet 2014).Erikson and Blanck would, however, respond with a thorough critique of subjective measurements of quality of life, ranging from a dismissal of Bentham's utilitarianism and its principle of happiness as the goal of government, to contemporary happiness studies and their revival of Benthamian styles of thought (for example Layard 2005).
As a response to the request for a new value meter by the Ministry of Finance, Erikson and Blanck presented a set of indicators for measuring quality of life and welfare that stretched back to the work of Sten Johansson.As presented above, Johansson had conducted the Level of Living survey, a large-scale inquiry on levels of welfare in which a set of indicators were in turn inspired by the United Nation's recommendations.Erikson and Blanck argued that Johansson's survey also had forward-looking qualities because 'the components that Johansson assembled in the Levels of Living survey of 1968 are principally the same as those found in the Stiglitz commission' (Erikson and Blanck 2015: 55, our translation).Nations 1961, p. 4, Johansson 1973, p. 214, Stiglitz, Sen, and Fitoussi 2009, p. 14-15, Erikson and Blanck 2015, p. 23.The concepts have been ordered with regards to the United Nations (1961) order.Numbers are the original order in each document.
However, what would become a matter of controversy was Erikson and Blanck's principal dismissal of 'subjective' indicators.Their arguments against such measurements in their report could be summarised thus: • Political decisions must be grounded in factual circumstances, not subjective experiences.
• Subjective experiences of welfare are prone to adapt according to factual circumstances.
• Using subjective indicators politically may lead to ignoring objective changes in welfare.
This critique entailed both facts and values.So-called subjective indicators were argued to be unstable for reporting back because of the hedonic adaptation in humans.However, more importantly, such measurements were seen as threatening the value of welfare because, as a 'factual' circumstance, the 'subjective' indicators would divert attention from the 'objective' changes in society.
Table 1 shows a side-by-side comparison of the various sets of indicators measuring welfare and quality of life noted by Erikson and Blanck, who argued that there exists a line of continuity between, on the one hand, the international indicators stated by the UN and the Stiglitz report, and on the other hand, a tradition of research called the 'Scandinavian model' of welfare research.This expressed line of continuity serves as a point of entry for identifying two historical threads that precede the contemporary debate on welfare measurements in Sweden.Lines of continuity, as Kuhn argued, are crucial to making the current research paradigm look linear and cumulative (Kuhn 1996: 138-40), and this also goes for value meters.If they appear as if they came from nowhere, each component has to be motivated and scrutinised, whereas if they build on previous findings and standards, they will appear as less controversial.
The indicator 'economic resources', as found both in Johansson's 1973 account as well as in Erikson and Blanck 2015, refers to Richard Titmuss's notion of 'command over resources', as discussed earlier.
Similarly, as the indicator 'political resources', it has a special focus on the individual and his/her 'objective' circumstances.The space of action, which is either widened or closed, should be the locus of both measurement and measures taken by the state.Defining these indicators rather as 'potentialities' or degrees of freedom also mark what the Scandinavian welfare researchers saw as the limit of state intervention.So-called subjective indicators, such as happiness or life satisfaction, would be 'final values' in such a perspective, and should be off limits for the state to decide upon.
The indicators proposed by the Scandinavian welfare researchers resonate well with the 2009 Stiglitz report, at least on one level.The 'capabilities approach' of Amartya Sen, one of the co-authors of the report, chimes closely with the 'command over resources' perspective advocated by Johansson and Erikson, and Blanck.However, the Stiglitz report acknowledges both subjective and objective indicators as central to the measurement of welfare.Stiglitz et al. (2009), with especial regard to life satisfaction and (feelings of) insecurity, explicitly point to the value of including 'subjective' measurements, something that Erikson and Blanck clearly reject in their interpretation of this important report.
As all Official Reports from the Swedish Government are subject to a round of referrals in which selected organisations and authorities review the report and return their verdict to the department that commissioned it, there is rich material for analysing opposing sides in knowledge production.Erikson and Blanck's report was dispatched to 76 consultation bodies (Gumpert 2015) of which 65 responded.In this section, we will give special attention to those referrals that concern the subjective/objective controversy.

Scientif ic suppor t f or subjective measurements
The most comprehensive critique of Erikson and Blanck's report came from a number of Swedish academic institutions.The universities of Stockholm, Gothenburg, Linköping and Umeå all stressed, in contrast to the report, that 'subjective measurements' should be considered and included in the definition of welfare.They argued that scientific inquiries had shown considerable progress in this line of inquiry and that there existed a strong international frontier of research.For example, the School of Business, Economics and Law in Gothenburg argued that 'there is an increasing and important international strand of research concerning happiness and subjective well-being, and plenty of valuable information can be extracted from this variable' (Handelshögskolan, Göteborgs universitet 2015).The Faculty of Humanities at Stockholm University asserted that ' [Erikson and Blanck's indicators] should be expanded to also include subjective measures of well-being and quality of life.This follows current research and international standards and such measurements are highly relevant for policy issues' (Stockholms universitet 2015a).Furthermore, the medical university Karolinska Institutet went even further and stated that 'The report advances objective measurements of quality of life-which seems strange, as the point of departure for quality of life can be seen as a subjective judgement' (Karolinska institutet 2015).Umeå University even pointed out two 'misunderstandings' in Erikson and Blanck's report, first that '[falsely] believing that life satisfaction and well-being (sometimes summarised as happiness) cannot be expected to change over time'; and the second that 'believing that there will be absurd consequences if governments act to maximise the subjective well-being of the population' (Umeå universitet 2015).
Thus, several universities referred to scientific advancements in measuring subjective indicators.But they also mentioned policy reports that had begun using such indicators, and seemed less concerned about the possibility of harmful political uses of subjective measurements.
However, support for subjective indicators was also strong among other institutions writing referrals to Erikson and Blanck's report.Statens beredning för medicinsk och social utvärdering (SBU) stated that they did not agree with the conclusions as to 'why the commission has chosen to take no account of subjective measurements when quality of life in itself is a subjective concept' (SBU 2015).Länsstyrelsen i Stockholms län pointed out that the report 'stood in contrast to the methodology adopted by many other countries and transnational organisations (for example, the UN Happiness Report)' (Länsstyrelsen_Stockholm 2015), a similar argument as in Folkbildningsrådet which noted that the subjective dimensions ought to be included just as in the OECD report (Folkbildningsrådet 2015), and the Institutet för Framtidsstudier claimed that 'international recommendations' strongly advised subjective indicators to be complementary to objective ones (Institutet för Framtidsstudier 2015).Even if the scientific issues concerning measurements were mentioned, these referring bodies primarily argued that Sweden should adopt 'subjective' measurements because they were already used and recommended internationally.

Scientif ic suppor t f or objective measurements
Although subjective indicators were widely supported among the consulted bodies, there were also actors that supported Erikson and Blanck's strict adherence to objective indicators.Uppsala University wrote in defence of the report's rejection of 'subjective' indicators: Concerning subjective measurements the commission's rejection is convincing, as it departs in that only measurements rendering visible changes over time shall be in question.It is shown how subjective measurements are far too dependent on levels of aspiration and adaptation to circumstances, making them unsuitable for comparisons among [social] groups.(Uppsala University 2015) As mentioned above, Stockholm University had argued in favour of subjective measurements.However, their referral was in fact written in a rather ambivalent fashion since two separate faculties were consulted (this is sometimes the case with large universities).While the Faculty of Humanities had advocated for the inclusion of subjective indicators, the Faculty of Social Sciences-home to the Swedish Institute for Social Research (SOFI) where Robert Erikson was still active-instead argued along opposite lines, by stating that subjective indicators were in fact rather useless: The faculty of social sciences would like to add that the values of subjective indicators of life satisfaction or happiness are strongly limited as they are not knowledgeable about actual life circumstances.Aspirations, discontent, and satisfaction are determined by actual circumstances and by circumstances relevant to the lived lives of the persons under investigation.Measurements of happiness are of diminutive value if reported exclusively.(Stockholm University 2015a) In these accounts, subjective indicators were seen as problematic; both from a scientific point of view due to the adaptation and aspiration effects and because they did not portray the 'actual' circumstances that in turn would affect people's perceived quality of life.This critique, once again, bears resemblance to the Easterlin paradox from the mid-1970s (Easterlin 1974) and Johansson's theoretical point of departure from the Level of Living surveys (Johansson 1973).
Since the SOFI, which was founded in 1972 and integrated with the Faculty of Social Sciences in 1994, had been the host of the Level of Living surveys for such a long time, it is not surprising that they dismissed subjective indicators.What is striking about the referral of Stockholm University is instead how the Faculty of Humanities clashed with the Faculty of Social Sciences in the same referral text, thus leaving a response full of contradictions.The Humanities faculty used words such as that they 'criticise strongly' the interpretation of quality of life as 'freedom of action', as advanced by Erikson and Blanck.Moreover, the Humanities faculty dismissed the idea that 'subjective' indicators cannot describe changes over time (as claimed by Erikson and Blanck) as 'obsolete' and 'false' (Stockholms universitet 2015a).
We followed this controversy a little further by requesting all additional material from Stockholm University.We discovered two preparatory works, which were the sources of the above referral.Here the Faculty of Social Sciences was more precisely defined as SOFI and the Department of Criminology (Stockholms Universitet 2015b), and the response by the Faculty of Humanities was authored by Henric Hertzman and Bengt Novén (Stockholms Universitet 2015c).In these preparatory works the controversy appears in a much clearer light.
The Faculty of Social Sciences wrote that they supported all aspects of Erikson and Blanck's report, except that Statistics Sweden (SCB) should be responsible for the collection of data.Instead they suggested that this responsibility should be delegated to academic researchers.They briefly listed a number of existing academic institutes as potential candidates for the mission, and also wrote that 'under the right circumstances, SOFI looks positively at the opportunity to act as the host institution' for such data collection (Stockholms Universitet 2015b).In other words, SOFI not only supported Erikson and Blanck.They also had a positive view on actually performing such measurements.This formulation was not, however, included in the official referral by Stockholm University, and consequently not forwarded to the Ministry of Finance.
The Faculty of Humanities, however, clearly appears as one of the fiercest critics of This way, the Faculty of Humanities reached all the way back to the problem that the Scandinavian welfare researchers faced in the mid-1970s, only to turn it upside down.The role of a new value meter of welfare should not, according to Novén and Hertzman, tell politicians what types of decisions to make; they should only report back facts free from 'ideology' and 'normative' values.In other words, the role of the expert was questioned, with regards to (the possibility of) neutrality in the measurements that were to be handed over to politicians.
The differing positions on subjective versus objective measurements in the referrals crack open a recurring controversy in the social sciences and their relation to policy indicators-or in other terminology the role of the social sciences in creating value meters.The contrasting ideals concern two interrelated questions: what can be measured accurately, and what should be measured when the results are handed over to policy makers?
Advocates of subjective indicators have struggled to deal with the effects of (hedonic) adaptation, and this is used as a counter-argument for implementing them as indicators.If people return to the same levels of happiness and life satisfaction, even if their objective circumstances change (for better or for worse), the measurements have no meaning for planning the welfare society.However, even though this discussion has been returned to since at least the 1970s, subjective measurements have been implemented as standard measurements throughout the social and medical sciences in numerous studies (see Kullenberg and Nelhans 2015), and they have become ubiquitous indicators in a number of international reports.

Discussion: Calibrating t he Value Meters of Society
Deciding on what to measure and, as a consequence, how to measure it does not proceed in a straightforward fashion from point of departure to a gold standard.As we have shown, in the case of welfare indicators, the process of deciding on indicators is a site of negotiation and controversy, where various concepts struggle to define what will be the future measurement of society.To account for the value of society, value meters (such as GDP, kilogram, stock market index, Gross National Happiness of the country of Bhutan, etc.) have to be invented and made durable, both in technical terms and as politically successful measures of some kind of value.When these devices work, in both terms, they make valuation (Helgesson and Muniesa 2013) of the welfare of nations possible.
With this article, we have tried to draw attention to the negotiations present in constructing such a 'value meter of welfare', a device that will by necessity involve a reduction of millions of interactions with statistics.To obtain data, the procedure must first have defined what data to look for.Thus, in this article we have examined how three episodes of debates on alternative valuation measurements to GDP have been related to broader international contexts, which entailed both scientific and policy-related elements.We have investigated how scientific arguments have been mobilised in debates on replacing GDP in favour of a new measurement of welfare and mapped out which actors have been involved in modulating various lines of argument in defining what values are worth counting.
We have shown that there are two interconnected sides of value meters-facts and values-that are fused together to form semi-stable devices that can be used to measure welfare.Methodologically these two sides, each having quite different modes of veridiction, can be difficult to bring into comparison.In this article we have deliberately chosen documents that yoke together facts and values.The way arguments are mobilised point on the one hand to the direction of establishing a measurable relation to something that can be detected with the value meter, such as housing, schooling or life satisfaction.Such knowledge can be put on trial as chains of reference, as social scientific knowledge.However, on the other hand, the value meter also has to be able to take measures in relation to values, such as creating a more 'sustainable' society (as the Green Party argued) or towards a greater 'equality' in society, as the Scandinavian welfare researchers kept coming back to.As these two modes of veridiction are conjoined, the process of firmly establishing the value meter can proceed; but correspondingly, the failure of either side may also bring about a crisis in measurement, as has been-many would argue-the case with GDP for almost half a century.
For a value meter to work, it needs to be backed up by a science.Just as GDP had to be accounted for with reference to economics (Halsey 1934), subjective measures of welfare have often been referred to as a new science of happiness (Layard 2005) in which measures of happiness, well-being, quality of life and life satisfaction have ended up in several important reports written in the first decade of the new millennium.However, in order to make 'subjective' forms of measurement credible enough to be put on equal footing with 'objective' indicators of GNP/GDP, economists, psychologists and social scientists had to start conducting empirical and quantitative studies that could compete with those supporting other value meters.This is why we have attempted to analyse how 'actors set up the collective socio-technical agencements that make valuation possible, stable, credible, accountable, and liable to compete with alternative perspectives on value' (Kjellberg and Mallard et al. 2013: 22).In part, the stability of the 'objective indicators' approach stems from the way sociologists were able to display a line of continuity going back to the United Nations indicators in the 1960s, and in particular to the Levels of Living surveys conducted more than 40 years ago.For the subjective approach to show the same historical lineage, there are no comparable results or devices to fall back on in the Swedish context.Instead, the advocates of subjective indicators had to start more or less from scratch at the turn of the millennium, as important research findings and reports, such as the Stiglitz commission, were mobilised.
However, scientific stability is not the only matter of consideration when constructing a value meter.The core set of researchers in the Scandinavian welfare research were all concerned with a sense of political stability centred round the interface between the market and the welfare state.Johansson's (1970) theoretical model of social integration especially was based around the political relationship between the labour unions and the employers.Higher levels of living were the outcome of a successful relationship between these actors, and these levels could be assessed 'objectively', Johansson argued.This stands in stark contrast to the 'subjective path' chosen by his contemporary Anglophone researchers, for example Hadley Cantril's influential The Pattern of Human Concerns (1965).Here, attention was drawn towards the subjective and immediate experience and the impact of that experience on behaviour.If people were satisfied with their life, unrelated to their actual life circumstances, they would express their aspirations and adjust their behaviour accordingly.In other words, stability could be achieved by monitoring the concerns of the people, and then adjusting politics based on that knowledge.This is what the Swedish sociologists clearly viewed as a form of political manipulation, and they have defended their position ever since.
However, the subjective turn in happiness research grew stronger, and this line of studies would have similar outlooks to Cantril's work.Cross-national surveys, which did not exist readily at hand for the economists of the 1970s, fuelled the trend of 'subjective' indicators, both as they provided large amounts of data, and would produce continuous indices where nations could be compared with each other.Happiness economics and new measures of welfare were constructed as a detour via scientific research, before it could be implemented in policy reports.Studies by Ed Diener (Diener et al. 1995(Diener et al. , 1999;;Diener 2000), Rut Veenhoven (Veenhoven 1991) and Ronald Inglehardt (Inglehart 1990) were based on large-scale cross-national surveys, such as the Eurobarometer (since 1973), the World Values Survey (since 1981) and the World Database of Happiness (since 1980).These surveys have often been cited not only by researchers, but also by the mass media, which frequently publishes ranking lists of 'the happiest countries'.Here, subjective indicators were implemented and calibrated, tested against vast amounts of survey data, and were then turned into indicators during the 2000s.
When value meters are made to work, facts produced by the social sciences have multiple ways of escaping their scientific contexts to become constitutive elements of state measurements; or to use a more familiar term: they may become statistics (Porter 1995;Desrosières 1998).The negotiation for determining indicators of welfare is a constant endeavour towards decreasing uncertainty and assembling a value meter that can be put to work outside the contingencies of scientific knowledge.It is primarily when value meters are discussed in relation to the state that the tension becomes intense.This explains why the group of sociologists behind the Levels of Living surveys could treat 'subjective measurements' as intellectually stimulating on a scientific level, but at the same time ward them off completely when official statistics were under consideration.
How welfare and quality of life is measured will affect how welfare is made.Social scientists play an integral and performative role in the co-production of scientific devices and social values as they create many important value meters that we live by in modern societies.However, new measurements are not easily invented and immediately adhered to.Rather, as we have tried to show in this historical account, they change only after series of negotiations, and sometimes fall back on measurements stabilised decades ago.
Episode 2: A Happiness Index? Green Par ty Proposals 2007-15 So-called subjective indicators as components of a measurement of welfare are of recent date in Swedish political discourse.Even if such 5 measurements had been presented internationally by academic researchers in the Social Indicators Movement since the mid-1970s, Political proposals to find alternative measures of welfare had already been put

Table 1 .
Comparison of indicators for measuring quality of life.English in original, see Erikson and Blanck.Novén and Hertzman, whosigned this preparatory work, connected the advances made in research on subjective indicators to international research and argued that Erikson and Blanck criticised subjective indicators on false and outdated grounds.Moreover, they drew from the Stiglitz report the same conclusion as the Green Party proposals, namely that it supported 'subjective' measurements.However, from the perspective of value meters, what is more interesting is the discussion on the relation to political values in measuring welfare.Novén and Hertzman wrote:We find it unsuitable to select measurements of quality of life, which depart in controversial normative presuppositions about the role of politics in an equal [socially just] society … Ideological considerations about what is a suitable decision for politicians to make should be left out of this discussion.(Stockholms Universitet 2015c)