Social work, emotion management and the transformation of the welfare state

Summary Drawing from the sociological and organizational scholarship of emotion management in human service work, our historical analysis elaborates on how external pressures and efficiency requirements jeopardize employees’ ability to carry out emotional labour in the public sector. Building on the Finnish social workers’ accounts published in a professional trade paper in 1975–2009, the article provides an insight into the emotional challenges and dissonances of interactive work. Findings The article suggests that social workers’ emotional job requirements are deeply embedded in and influenced by the broader social context and changes in the welfare state. The article indicates how workers’ emotion management underwent changes and readjustments from the heyday of the expansive welfare state of the mid-1970s, through the severe economic recession of the 1990s, and was bound by the ideological reform of the Finnish welfare policy in the 1990s and 2000s. Applications Historical prospect to emotion management shows social workers, social work students and the academic community how employees’ emotion management has been gradually recognized as a crucial element of relational client work over the decades, but also how workers’ emotional job requirements and demeanours are intertwined with broader societal changes. Second, emotion management research provides a solid framework to elaborate emotional and ethical dissonances which are embedded in social work more profoundly.


Introduction
In current working life, emotional labour is highly prevalent in human service work. More than 30% of European employees frequently engage in emotional labour in their work (Eurofound, 2016). Several female-dominated occupations in the public sector, such as social workers, nurses, teachers and psychologists are often characterized by demanding emotional interactions which require emotional skills to handle the challenges arising in their labour processes. Numerous studies have acknowledged that social workers suffer from occupational distress and are more vulnerable to emotional exhaustion and depression (Blomberg et al., 2015;M€ antt€ ari-van der Kuip, 2015;Travis et al., 2016). A recent study shows that the risk of disability or early retirement due to mental diagnoses amongst social workers is twice that amongst psychologists, special education teachers and kindergarten teachers in Sweden and Finland (Rantonen et al., 2017). Sociological analyses have provided important insights into the characteristics and nature of emotional labour in human service professions in general (Ashforth & Humphrey, 1993;Grandey et al., 2013;James, 1992;Mann, 2004), as well as into how social workers experience and cope with emotional labour (Gorman, 2000;Gray, 2012;Moesby-Jensen & Nielsen, 2015).
The purpose of the article is to investigate social workers' emotional organization of work with a particular emphasis on how the underlying historical and structural aspects have redefined the emotional labour processes of social work. Therefore, our qualitative analysis is reinforced and driven by a historical and sociological approach. The analysis and the outcomes of the article are based on the social workers' written contributions published in their official trade paper, covering the period 1975 to 2009. As is typical in union publications, the employees and their representatives pay attention to the wage and status of work, lobbying, working conditions and an affirmation of professional identity. However, they also discuss and debate the burdensome aspects of work, the exigencies of interactive jobs and work exhaustion.
The underpinnings of this study originate from sociological and organizational studies on emotional labour and emotion management (Grandey et al., 2013;Hochschild, 2003;Lively & Weed, 2014;Wharton, 2009). Grandey et al. (2013) identify three different approaches that characterize the main trends of emotional labour research: (a) occupational requirements, which are explored in sociological research, (b) emotional displays in organizational studies and (c) intrapsychic processes in the field of psychologically oriented research. These approaches have gradually differentiated into separate fields and Grandey et al. (2013, p. 17) have underlined the importance of exploring the 'dynamic interactions' of these disciplinary approaches. Intrapsychic emotional regulation is extensively studied in occupational and organizational psychology (Grandey et al., 2013) but this approach tends to be limited to 'inside' micro-environments, to subjects such as supervisor and social support and organizational climate. Given this, this article draws attention to how social workers' emotion management is incorporated in the broader context of interactive work.

Modes of emotion management of human service professionals
Emotion management and professional feeling rules are focal conceptual lenses for exploring the intra-and intersubjective dynamics of employees and organizations. These conceptions draw on Arlie Russell Hochschild's (2003) book The Managed Heart. However, Hochschild's contribution generated a vivid debate on whether the concept of emotional labour was an adequate lens for exploring such highly educated professionals as nurses, teachers and social workers. Hochschild notes that these professional groupings engage emotional labour in their mundane work though their work is not directly controlled by leaders and their emotional displays are not driven by profit. Moreover, Wouters (1989, p. 100) remarks in his critical reading of The Managed Hearth that professionals who supervise their own emotional labour actually 'put stronger demands on their emotion management'. Consequently, one group of scholars prefer to use the concept of 'emotion management' (Bolton, 2005;Lively & Weed, 2014;Morris & Feldman, 1997) as an umbrella term, in order to avoid the baggage of the 'commercialization of feelings' which leads to the negatively coloured notions of emotional labour.
This article defines emotion management as a self-regulatory process in which employees control subjective emotional states, emotional dispositions and organizational behaviour. The self-regulatory process is guided by both formally and informally internalized professional feeling rules to achieve organizationally desired goals. We prefer the term 'feeling rules' instead of 'display rules' because its focus is on human service workers' experiences and self-perceptions of their emotion management. Professional feeling rules are internalized via professional training and occupational socialization (Ashforth & Humphrey, 1993;Rafaeli & Sutton, 1987), but employees have some degree of autonomy to self-regulate, adapt and resist the emotional scripts of work (Bolton, 2005;Erickson & Stacey, 2013). Employees' emotion management 'both reflects and influences' occupational requirements and organizational contexts (Bolton, 2005;Erickson & Stacey, 2013, p. 177). Ashforth and Humphrey (1993), amongst others, state that Hochschild overplayed the organizations' power to manipulate labourers' emotions. They emphasize the normalcy of displaying 'genuine' feelings at work, leading to positive outcomes of emotional labour. This argument has advanced the research on human service work (Bolton, 2005;Erickson & Stacey, 2013). Erickson and Stacey (2013) claim that the definitions of commercial emotional labour are inadequate for understanding the work of human service and caring professions. They highlight the incommensurability of the emotion practices of the 'caring' professions compared with customer service work. Overall, human service professionals are responsible for their clients and patients and are guided by the ethical principles of their profession rather than by sheer commercial profit. Bolton (2005), who introduced a versatile typology of emotion management at work inspired by her research on nursing, further elaborates on many of these critical insights. The typology stresses professional identity formation at work, the multiplicity of emotional scripts in organizations and employees' agency in emotion management. Bolton differentiates between pecuniary, prescriptive, presentational and philanthropic emotion management.
Pecuniary emotion management stems from Hochschild's original theorization on commercial frontline service work, whereas prescriptive emotion management is determined by the features of professional feeling rules which are typical of public and caring professions and their organizational emotionality. Bolton (2005, p. 95) remarks that prescriptive emotion management is dictated via 'membership in a professional body' and motivated by social recognition and status. Prescriptive and philanthropic emotion management are neither directly or strictly managerially controlled, unlike pecuniary emotion management, and therefore these modes of emotion management contain more occupational autonomy and allow more room for emotional scripts in client interaction and workplace sociability. Presentational emotion management denotes the general social interaction rules that are internalized in primary socialization processes. In turn, philanthropic emotion management refers to organizational emotionality, which embraces supportiveness, congeniality and compassion, and is motivated by 'general good' and a moral belief of going the 'extra mile' in social exchange (Bolton, 2005;Lewis, 2005).
Nevertheless, although Bolton's typology and the conceptions of emotionality and emotion management of human service work (Erickson & Stacey, 2013;Mann, 2004) have proven to be novel insights into emotional organizations in human service work, the historical dimension remains an overlooked element of interactive work (Anttila et al., 2017). Secondly, the history-sensitive approach underscores the refinement of the occupational dissonances embedded in institutionalized public sector work, which is largely carried out by female-majority occupational groupings.

Social work in the Nordic welfare state
The objective of the article is to expand the understanding of emotion management in social work in a specific historical period  and in a specific national context, Finland The Finnish welfare state can be defined according to Esping-Andersen's (1990) typology which categorizes welfare regimes into liberal (United Kingdom), conservative/corporatist (Germany) and social democratic (Sweden). These classification labels do not refer to political ideologies as such but indicate the patterns of responsibilities of welfare institutions and the focal providers of social services. The Finnish welfare state epitomizes the social democratic, institutional and service-oriented 'Nordic model' due to its high proportion of welfare costs of gross national product (GNP; Kautto et al., 2001). In 1970, welfare costs were 13.6 per cent of GNP, and increased to 25.7 per cent in 1990 (Kosonen, 1998, p. 110).
Symptomatically of the institutionalized phase of the Finnish welfare state in the 1970s, public welfare services provided expanding employment opportunities, especially for educated women (Anttonen, 2002;Satka, 1995). In addition to the municipal service sector and child protection, social workers were hired progressively more often in state schools, health care, rehabilitation and correctional services (Vuorikoski, 1999). Our period of analysis tackles an era that begins in the heyday of the 'universalist' welfare state of the mid-1970s, and continues through the severe economic recession of the 1990s, which was soon followed by a social recession and the ideological reform of Finnish welfare policy (Kettunen, 2006;Timonen, 2003). Despite being a latecomer, the Finnish welfare state caught up with its Scandinavian counterparts in the 1980s in terms of its level of welfare institutions and services (Kangas & Palme, 2005;Kuhnle & Alestalo, 2000). Notably, this occurred concurrently with the welfare state reforms and retrenchments that swept over many European nation states.
The article suggests that changes in welfare policy and ideology, not to mention economic upheavals, have influenced and reformed the prerequisites of emotion management in social work. Our approach builds upon the historical perspective and its unique prospect for investigating the temporal changes and trajectories that define emotional organization through an analysis of social workers' recollections. Analysing social workers' perceptions of emotion management over several decades enables us to explore the contingent socio-cultural factors that determine their emotional job requirements. By drawing on human service workers' accounts, the article complements and elaborates on the research on work sociology and emotion management, which regularly focuses on the workers of liberal welfare states.

Method
The researchers' decisions in relation to the research material and the theoretical framework for the article both involve and contribute to the conclusions (Kalela, 2012). Hence, it is important to briefly clarify these decisions. The solidity of the findings and the outcomes of the article consist of three phases: (1) the selection of the relevant source to meet the research objectives, (2) the selection of the social workers' written contributions to analysis and (3) an in-depth analysis of these contributions. First, we selected the nation-wide trade paper of the social workers as the main data source because it is the official organ of the Union of Professional Social Workers. The journal is a noteworthy forum for professional discussions on the policy and practice of social work. The union membership rate was over 80% during the period selected for analysis and the journal was delivered to all its members. In 1975, the journal was published under the title Sosiaality€ ontekij€ a ('Social Worker'). The social workers' trade union changed its name in 2002, and in the process the trade paper was renamed Talentia (indicating the group of skills and virtues). The material varied greatly in terms of length and contents, from paragraphs to full-length articles, and included editorials, columns, complete articles and interviews with social workers, as well as letters to the editor.
In the second phase, we carefully collected social workers' contributions for further in-depth analysis. We went through all the volumes of the trade journal from 1975 to 2009 and chose 355 articles that we found the most relevant and representative in terms of the emotional organization of social work. As the contributions were written over a long period of time, we were able to identify the subjects that raised considerable attention amongst social workers during these years. This research phase implies that the struggle with workplace emotions was an acute issue for social workers. Social workers extensively discussed client work, controversial organizational expectations as well as job autonomy, job requirements and the emotional costs of work.
In the third phase of analysis, we carried out a responsive re-reading of the material to trace the temporal changes of the social workers' emotional responses under study. We focused on how intensive and extensive the social workers' accounts were in order to recognize the amplitude and significance of the underscored matters. We paid attention to the changes in the debates and expressions that the social workers used to define the emotional demeanours and broader ramifications of work. We tracked down the issues that the employees considered as having an impact on their interactive work.
Our approach is best described as a theoretically informed analysis which strongly takes into the account both the historical context of the research material and, if possible, the occupational position from which the article is written. It is reasonable to assume that trade magazines not only constitute an agenda for trade unionism and social work activism but also give a voice to social workers to affirm their observations and communicate experiences. We view the documents as collectively shared manifestations as well as the negotiations of workplace emotions in social work in 1975-2009. We understand the social workers statements and responses describing emotional job requirements as the subjective perceptions of mostly frontline social workers. Although they capture the experiences and feelings of the workers, as well as the topical debates of the period, they also entail ideological statements and intentions to influence both union leaders and policy-makers.

Results
'Warm yet strong-minded' -A social worker in the era of the institutionalized welfare state in the mid-1970s and 1980s In the mid-1970s and the 1980s, many social workers, trade union leaders and social work teachers talked extensively about the emotive nature and emotional demands of social work. This debate was rooted in the noteworthy aspects of social work, such as client work and the emotional demeanours of social work. Leading scholars had already addressed the emotional interactions of social work in the formative psychodynamic years of public social work in the 1950s (Satka, 1995), but in the following decades, it was the frontline workers who engaged in the discussions on the changing emotive aspects of social work.
The support and supervision of clients' social, psychological and economic problems necessitated social workers' emotional assessment, an empathic predisposition, and the management of interpersonal difficulties. Social workers' accounts and the union's committee reports stressed that their role as practitioners of institutional social policy did not make them workers therapists, and remarked that guidance, help and support were indispensable aspects of their work when managing client's multidimensional problems. These notions reflected the ongoing development process in the welfare policy, which was marked by a strong desire to reduce economic and social inequalities in the 1970s and 1980s (Kautto et al., 2001). Social work in the era of the institutional and expanding welfare state from mid-1970 was fuelled in Bolton's terms by the principle of prescriptive emotion management accompanied with philanthropic ideals. Many leading social policy commentators were convinced that the characteristics of social problems were on the path to change due to the growth in material wellbeing and the gradual disappearance of absolute poverty. As a result, the prevention and management of citizens' 'psychosocial' problems become a profound objective of social work. Finnish reformist psychiatrist Claes Andersson encapsulates the problem of interpersonal sensitivity in his 1976 speech to social workers entitled 'Difficult client relationships': 'Andersson stated that helping another human being professionally requires a great deal of the helper. The helper's only tool is their own person, which they try to use to help someone else -the one who needs help. It is important to take a phenomenological stance. This means that the social worker dares to be open and unprejudiced, and comprehensively perceives the distress of another human being. It is important that a worker with empathy and the right orientation settles in the dialogue with the client'. (Social Worker, 6/1976, p. 3) As the citation indicates, the prevailing idea amongst social workers was that relationship building mattered and clients should be treated with respect. The emerging vocabulary in the trade paper advocated this development, as the impersonal and democratic-sounding term 'client' replaced such stigmatizing words, such as the 'poor' and 'needy' in the 1970s. The emergence of egalitarian forms of social exchange in social services was connected to the idea of the client's self-determination. These accounts reflect a broader informalization and democratization of social relations and the lowering of social hierarchies in the post-Second World War era in society more generally (Wouters, 2007). Hence, the democratization of the worker-client relationship reinforced the emotional intensity and attentiveness of social work, managing negative emotions, and ultimately social workers' significant responsibility for their clients.
Historically, however, Finnish social work has a strong legal-administrative tradition (Satka, 1995) which emphasizes the social worker's role as a civil servant.
Emotion management was at the crux of the social worker's contradictory roles as an official or supporter. In public social work, prescriptive emotion management was largely grounded on the regulations of the welfare policy and practice. The chairman of the union outlines this tension as follows: 'The role of a social worker as a municipal official is problematic. In order to coordinate the official regulations and guidelines and the individual needs and wishes of the client, social workers have to resolve many tricky conflicts of interest' (Social Worker, 6/1985, p. 4).
One of the most important skills in social work was the ability to 'read' clients and their emotional states on the basis of their gestures and bodily expressions. In addition to 'reading', social workers utilized metaphors such as 'listening' to describe their engagement with their clients. One social worker's quote of a client's opinion of a good social worker in 1987 was in line with this: 'Of course, one who listens too. Who can ask in a way that you feel that you can easily talk. It is a very delicate thing, a matter of subtle nuances. And then that transferring money is not the main thing' (Social Worker, 4/1987, p. 5).
Social work history researchers identify an analogous role ambiguity and carecontrol debate in the various national contexts and historical settings of social work (Hauss & Schulte, 2009;Jones, 2014;Lymbery, 2005;Reamer, 1998). Numerous social workers objected to the fact that the organization of social work was overtly determined by administration and that the historical past of bureaucratic tradition 'still exist(ed) in unwritten procedures' (Social Worker, 5/ 1986, p. 8). Though social workers were highly aware that social work was a regulated profession and justified by legal powers, they stated that professionalism did not necessitate facing the client in a bureaucratic or detached manner. To some extent, these notions were social workers' counteractions to the transnational academic debate on the so-called oppressive and bureaucratic social work in the mid-1980s (Baines & van den Broek, 2017;Jones, 2014;Lipsky, 1980;Reamer, 1998). These controversies demonstrate the challenges of emotion management in social work. However, the discussions of the trade journal also reveal how social workers actively engaged in negotiations on the aims and objects of social work.
Treating clients with respect and sympathy was an important professional feeling rule in social work. These feeling rules highlight how workers' perceptions of their professional identity and their occupational role were deeply connected with inter-and intrasubjective emotion regulation: 'What other occupation than social work provokes such intense emotions? Workers face clients who have acute needs or prolonged and multiple problems which afflict employees' resilience' (Social Worker, 3/1985, p. 16). During this period, social workers began to understand their core work through similar expressions and ideas to those that researchers used to define emotional labour in human service work in the next decade. These notions reinforced social work as an emotionally demanding human service work, comprising of similarities to nursing and teaching. However, social workers also considered the specificity and characteristics of their interactive work in relation to other welfare occupations, because social policy regulations constituted their focal prescriptive job requirements.
The struggle with gender. Social work has historically been considered a woman's work (Satka, 1995). However, the notion of gendered occupation was to some extent troublesome for social workers and their advocates, even though this fact was obvious to the practitioners of the 1970s . The commentators in the trade paper emphasized that men usually occupied managerial positions in social work institutions. Consequently, social workers gave ambivalent and differentiated accounts of whether social work itself operated a gendered institution and whether gender itself had an impact on emotion management. Namely, union-minded social workers and shop stewards questioned whether female gender entailed the inherent qualities for social work and its emotional job requirements.
According to some observations, a socialization process guided or pushed female employees into human service professions such as social work. Some commentators pointed out that females hold inherent characteristics which augment their sense to 'compassion, closeness and the responsibility of their fellows' (Social Worker, 5/1986, p. 10). However, social work activism addressed the conflicting outcomes of feminine inclusion in human service work; the responsibility for vulnerable groups in society, but poorly rewarded efforts. In 1986, one social worker remarked that the profession suffered from the occupational hierarchies of society, as employees were in a 'weak strategic position into the division of labour' (Social Worker, 1/1986, p. 18). This commentator also considered that working with the most vulnerable citizens did not improve either social recognition or occupational status in general. A great number of social workers saw the pitfall as springing from an unequal division of power and occupational hierarchies in the labour market in addition to underdeveloped professionalism, rather than from gender as such. Anttonen (2002) points out that social worker education and professional projects supported gender neutrality and obscured the gender issue in social work from the 1970s. We interpret that this agenda involved the closely linked professionalization of social work and efforts to improve occupational status and remuneration. Despite the academization of the profession in the 1980s, social work remained highly feminized. For instance, in the 1990s, trade union representatives noted that 90% of social workers were female. Therefore, social work emblemized a high level of gender segregation and the emotional division of labour in society. Still in the 2000s, social work activism strongly linked gender segmentation to the unfair reward of work which, according to commentators, had a negative impact on employee satisfaction and on workers' ability to conduct relational work effectively.

Impact of the economic upheaval in the 1990s
The Finnish welfare state and society faced a serious economic downturn in the early 1990s. The recession was one of the most severe on the European scale (Kosonen, 1998). It triggered a welfare reform that embraced the doctrines of market liberalism and New Public Management. The ideological restructuring towards the 'competitive state' (Kettunen, 2006) in the 1990s and 2000s in Finland has parallels with European welfare reforms which had taken place in the 1970s and 1980s (Carey, 2009). However, in terms of welfare institutions, the change was much more fundamental and far-reaching in many liberal welfare states (Baines & van den Broek, 2017) than in Finland. For instance, in the UK, the 'welfare state approach' was replaced by radical privatization and the 'community care reform' in the 1980s and 1990s (Harris, 1998;Healy, 2009).
The debates in the journal demonstrate that social workers were witnesses to socio-economic turbulence and the distress of vulnerable citizens. The economic restructuring in the first half of the 1990s not only amplified the number of service users, but social workers stated that their clients suffered from even more serious financial, social and psychological trouble. From the social workers' point of view, time pressure and the increased workload left their mark on job discretion and thus provoked expressions of cynicism and feelings of helplessness. Social workers' emotional responses became entangled with these extensive quantitative changes. According to a report of the National Institute of Health and Welfare, in 1990, 180,000 families were subsidized by income support, whereas in 1995 this figure was 330,000. During this time, the number of children taken into care increased by 17% (Social Worker, 3/1996, p. 16). In addition, in the first half of the 1990s, the real income of social workers decreased; the proportion of unemployed social workers increased, and the number of social work posts began to decrease, which was unheard of in the history of the institutionalized welfare state (Vuorikoski, 1999).
The articles in the trade journal in the early 1990s point out that social workers found that extra-organizational changes were excessively governing the terms of social work. They forcibly opposed the predominant policy: Creating trust in the client and acting as their advocate is essential while calculating the sufficiency of resources. You must be empathetically economical. You have to make individual decisions and at the same browse laws and regulations. We must strive for integrity and at the same time write appropriate reports for the accountant. So, as my supervisor said, this work requires the character of a masochist. (Social Worker, 5/1992, p. 23) The quote above underlines how social workers witnessed the arrival of workplace emotions that resembled pecuniary emotions. It also reflects the workers' feelings of helplessness as they thought that they had a very limited autonomy to resist these retrenchments. Many commentators pointed out that the welfare cuts and administrative reforms in this period consisted of disconnected objectives, such as organizational accountability and resource limitations, and the doctrines of clients' 'activation' and 'empowerment', which were underpinned by an effort towards organizational efficiency and the hope of reducing welfare costs. The quote above captures the employees' feelings that their professional discretion and emotional regulation were constrained due to extra-organizational causes. The decisive implication of readjustment and economic downturn led to accountability at the level of both bureau and individual employee.
In 1996, a social worker formulated a list of job demeanours that exemplified the pressure on employees' to adopt self-focused emotion regulation, aiming for a desired organizational behaviour to effectively influence clients: • 'a smile is a vital part of a social worker's outfit; smile when it is possible and smile in a such way that it can be heard • maintain eye contact with the client all the time when speaking to them • always try to put yourself in the client's position • remember that your looks reflect your image and attitude • make sure that you are aware of the client's situation; if you do not know the answer, obtain the required information • you can make the client's service memorable by showing a genuine interest in them • strive to make the client feel good so that they do not forget you, and the next time the meeting will not be so frightening' (Social Worker, 3/1996, p. 13) The social worker wished to expand the list on the basis of her own professional experiences: • 'do not settle for banalities • do not hide behind being in a hurry • remember that the work community is like a team whose competence is dependent on everyone's commitment, co-operation and a mutual will to give their best' (Social Worker, 3/1996, p. 13) This example points out the refinement of social workers' emotion management compared with the late 1970s and 1980s notions of client work. Notably, the abovementioned requirements resemble pecuniary emotion management, the primary goals of which are customer satisfaction and service-orientation. In a sense, the statements remodify prescriptive feeling rules by suggesting that the social worker should possess the qualifications of commercial service industries for better organizational and professional performance and task effectivity. The similarities to 'aesthetic labour' and feminine 'soft skills' (Witz et al., 2003) are striking. Interestingly, the social worker who published this list in the trade paper revealed that these instructions rephrased Disney World's official letter to its workers. Job demeanours such as one's 'look' and smiling are more likely to be found in corporate manuals for frontline customer work typical of the 'smiling factories' of Disney World (Van Maanen, 1990). Although we do not argue that the differences between commercial and human service emotion management emerged fully in the latter half of the 1990s, our findings reveal the blurring boundaries of workplace emotions, mainly due to the implementation of efficiency requirements and the reaffirmation of customer orientation. In a nutshell, the client was portrayed as a 'citizen-consumer'.

Challenges of emotion management in the era of retrenchment in the mid-1990s and 2000s
It is emblematic of the era of retrenchments that employees felt that significant emotional ambivalences originated from the inconsistencies between the demands for organizational effectiveness and the professional feeling rules backed by ethical standards. Another, and arguably conjoined significant issue was social workers' ability to manage the provocative emotional responses to their work. During these years, serious encounters, occupational violence and agitation management gained increased visibility amongst social workers.
Social workers identified emotions as an intrinsic part of their relational work, but during the 1990s and 2000s, affect and self-focused mood management especially troubled workers. One social worker encapsulated the emotional desolation of social work in 1998 by headlining their article: 'When the only tool is yourself' (Social Worker, 9/1998, p. 11). Based on their training and occupational experience, social workers were mostly aware of the risks of over-identifying with the client's troubles. Many social workers felt that clients with difficult life situations played on their minds and caused recovery difficulties. Numerous social workers felt that child custody decisions were the most painful part of their work. Regardless of the differences between professions, social workers' difficulties in containing their emotions have similarities with, for instance, those of nurses (Bolton, 2005;Lewis, 2005). Some social workers' contributions imply that they still remembered disturbing cases years later, which exposed them to long-term emotional suppression. Moesby-Jensen and Nielsen (2015, p. 698) found in their ethnographic study of Danish social workers that difficult cases 'got under their skin' and dominated their thoughts.
A social work undergraduate frankly expressed his thoughts on the practical training period in the journal in 1998, and made visible his personal struggle with crucial professional feeling rules; sympathizing with the client and non-judgmental attitude: Even during my training, my favourite clients and dislikes stuck out. My favourites are often single parents, young couples or anyone who tries to keep their affairs in order despite the difficulties. On the other hand, the pet hates of trainees are the young mournful lazy ones, who appear to be fit for work. Another pet hate is the alkie, who with no appointment insists on getting money because their previous social benefit has already run out. At times, it feels horrible to subsidize your pet hate and at the same turn your back on your favourite empty-handed client who you feel is more entitled to it for being so nice. Therefore, I find myself unloading guilt to another client, and keeping another in suspense as to whether I'll grant them income support. (Social Worker, 6/1998, pp. 7-9) These feelings may be defined as the reality shock of the newcomer, as well as a form of resistance and organizational misbehaviour. The trainee puts into words a focal problem observed in emotion management research, the suppression of genuine reactions and the inconsistency between job requirements and internal feelings. The trainee emphasized his difficulties in masking his inner reactions and displaying ambivalent responses to prescriptive and philanthropic emotion management. Parallel to cognitive dissonance (Festinger, 1957), in emotional dissonance the inconsistency of feelings is intolerable (Mann, 2004). Wharton (2009) discovered that putting on an emotional display on the job which is in conflict with one's personal feelings is positively linked to a higher level of burnout and emotional exhaustion. In addition to cynical and detached emotional orientation, emotional dissonance can enable understanding of the ongoing vicissitudes of the social work, the high level of employee turnover, trouble hiring skilled workers and escapism.
Social workers reported occasions when they felt that clients' abusive behaviour was a strategy for undermining their professional judgement and integrity to the advantage of the client. A social worker remembered the words of a lecturer who stated that 'clients can smell if someone is angry' (Social Worker, 4/1996, p. 16).
Violations of the rules of interaction bring out dark emotions. Tuija Virkki (2008) discovered in an interview study of Finnish social workers that if the rules of interaction were insulted, the acts of insults not only provoked anger and fear amongst workers, but also shame. Erickson and Ritter (2001) observed in their survey on gendered service sector employees that managing feelings of agitation, such as anger, irritation and frustration have negative health effects.
According to textbooks, there are no difficult clients. If a worker wishes to characterize a client as awkward, this sentiment is just a sign of organizational incompetence to meet clients' needs. These lessons come to my mind when a handgun shows up at the door followed by a client. Yes, I recall method lessons and according to them, with threatening clients we should create a relaxing, warm conversation, and aggression should be facilitated in a solution-oriented way. (Talentia, 7/2002, p. 8) The dissonance between the idealistic guiding principle and real-life encounters felt absurd yet threatening. 'Aggression' was merely an academic term in the textbook, but violence constituted an acute threat to the workers. One columnist questioned the instructions of social work education, which suggested that good interpersonal skills pre-empted social conflicts such as incidences with aggressive clients. Similar severe incidents that the social workers brought forth in their contributions questioned whether occupational safety officials and workplace practices were insufficient to guarantee employees' health and safety at work. In the mid-2000s, according to the survey, the union estimated that 20% of social workers had been subjected to violence. Mann (2004) found in her research on nurses that emotion management was consolidated mainly by an informal socialization process rather than by official means. Social workers' accounts reaffirm the significance of 'hidden' forms of emotion management, and then insufficiency of formal training to manage provocative emotional responses which typically originate from workplace violence, critical confrontations and experiences of emotional dissonance. Overall, the 'dark side' of emotional labour was not merely an abnormality or a dysfunction of organizational behaviour, but an integral part of human service work (Ward & McMurray, 2016).

Conclusions
Our analysis provides insights into the historically contingent nature of emotion management in social work. The article draws on emotion management research and elaborates on it by further incorporating broader historical and social perspectives. It aims to fill a gap between the micro and macro approaches of emotion management (Grandey et al., 2013) by illustrating how social workers' emotional responses incorporate extra-organizational attributes. In such a policy-driven profession, the changes in the welfare state created a feedback loop that influenced employees' intrapsychic management. This embeds emotional labour in a broader historical context; the welfare state, welfare reforms and economic turbulences. Erickson and Stacey (2013) claim that emotion management and employees' emotional regulation are not only shaped and governed by interaction rituals; they also reflect and involve larger social hierarchies and labour market segmentation.
The outcomes of this article show that despite the Nordic welfare regime containing unique features such as extensive public welfare services and a mainly highly educated workforce, the social workers' written documents reveal that social work is vulnerable to economic turbulences and bound by the international reforms of the liberal welfare states. The article contributes to emotion management research, but its approach also has considerable relevance for social workers, social work education and the management of social work organizations, as it draws attention to the core internal process of social work and the costs of emotional work for employees.
Social workers' perceptions are invaluable for shedding light on former quantitative observations of how human service workers are vulnerable to harmful negative emotions and emotional dissonance. On the other hand, the positive aspects of social work are also linked to relationship building and professional discretion, which necessitate the management of both the client's and the employee's affects. Interactive work and emotional assessments were recognized as the key features of social work, increasingly from the mid-1970s, and frontline workers gradually became more aware of the emotional cost and the 'dark side' (Ward & McMurray, 2016) of emotion management. Our analysis underscores the significance of the informal domain of emotion management (Lewis, 2005); these informal spaces facilitate sharing emotional skills amongst social workers and are invaluable due their supportive role in agitation management.
Another focal point of our analysis is that conflicting job requirements lie at the crux of social workers' emotion management. In sum, emotion management is a chief internal process for exercising professional discretion. Therefore, researching emotion management can deepen our understanding of social workers' role ambiguity and the multiplicity of their tasks (Jones, 2014;Postle, 2002). As the previous sections show, these modes of emotion management have diverse compounds and face many challenges in our specific historical period. Social workers struggle with the tensions between the ideals of philanthropic emotion management and the shifting welfare regulations and realities. In this regard, social workers have had to shape, reflect and readjust their own emotional labour processes in accordance with public welfare regulations. Bolton (2005) and Lewis (2005) show that employees use various types of emotion management and move between the versatile scripts of workplace emotions. This observation reverberates with our analysis, although our investigation emphasizes the role of the broader institutional logics of welfare policies and socio-economic changes in terms of the emotional organization of social work (Copp, 1998). From the mid-1970s to early 1990, the era of the expansive welfare state and the prevailing modes of emotion management were philanthropic and prescriptive. In the first half of the 1990s, the large-scale welfare cuts and the retrenchments driven by the economic downturn left their mark on the policy and practice of social work. This led to a sharp contrast between structural and institutional resources and the actual needs emerging from client work. The economic downturn was accompanied by the transnational influence of New Public Management, triggered by welfare retrenchments which echoed pecuniary workplace emotions due to the implementation of the cost-effectiveness ideas in social work. As an indication of the change in organizational atmosphere, an interview study showed that in the 1990s, social workers had nostalgic feelings towards the era of an expansive welfare state in the 1980s (Mutka, 1998).
Research material points out that an ideal social worker is adaptable, responsive, and devoted to managing interpersonal relations. Social workers and their advocates make it clear that their job requires high emotional skills but they hold an ambivalent stance on whether the attributes of gender contribute to their job demeanours. The feminine stereotypes of nurturing and caring women are associated with an 'innate' relational orientation towards identifying clients' expectations and their emotional states (Hochschild, 2003;Wharton, 2009). Drawing on a diverse set of research, Lively (2013) points out that female employees are generally more responsible for emotional balancing in interactive work. Hence, social workers struggled with the paradoxes of feminization at work. However, the prevailing discourse of social workers' professionalization, accompanied by the struggle to improve their occupational status, 'neutralized' the significance of gender from the 1970s onwards . Even though employment opportunities increased and social workers' educational requirements grew, union-minded social workers underlined the fact that gender distortion in social work was linked to their 'people' work skills being overlooked and poorly remunerated.
We have highlighted the broader context of emotion management in social work, and thus how the institutionalization and expansion of social work in Finland epitomizes larger trends in working life. According to the European Working Conditions Survey (Eurofound, 2016), about one in three workers in Europe report that they work in emotionally disturbing situations for onequarter or more of their working time. Emotional demands are more frequent in jobs that involve relational work. Recent large-scale register studies examining the current risks related to mental health between occupations have found that people working in occupations that involve human service are at a higher risk of sick leaves due to depression (Kokkinen et al., 2018) than those in occupations involving less human service. Amongst human service occupations, social workers and other groups whose work involves emotionally demanding situations are at a particularly high risk. Given this evidence, it seems likely that the emotional labour of social work captures the workplace emotions of a wide variety of occupations employed in current working life.

Limitations of the study
The main limitation of the research is connected with the extensive research material and its temporal coverage. Therefore, the article does not provide detailed analysis on the impact of welfare legislation on the labour process of social work.

Ethics
The study is approved by The Ethics Committee of Finnish Institute of Occupational Health.