What Makes an Ideal Unemployed Person? Values and Norms Encapsulated in a Computerized Profiling Tool[1]

Karolina Sztandar-Sztanderska, University of Warsaw

Marianna Zieleńska, University of Warsaw

1           Introduction

According to a growing body of literature, the payment of social contributions or the fact of belonging to a broad target group are no longer sufficient reasons to access services or obtain previously universal entitlements (e.g. Berkel & Valkenburg, 2007; Serrano Pascual & Magnusson, 2007; Dubois, 2009, 2014; Betzelt & Bothfeld, 2011). It marks a shift in the understanding of social citizenship which is no longer based on rights deriving from collective statuses, but instead, accessing benefits and services is becoming increasingly dependent on the assessment of individual behaviors and work attitudes. These changes create a necessity to find ways to assess individuals and compare them, and to reduce complexity of their specific circumstances in order to decide what type of public intervention or sanction shall be implemented. Previous research has shown that this kind of individualization in the frame of labor market policies, paradoxically, pushes for standardization: standardized tools are required to translate “differentiated life-situations into ‘manageable’ organizational categories,” (Garsten, Jacobsson, & Sztandar-Sztanderska, 2016, p. 286; see also Caswell, Marston, & Larsen, 2010; Garsten & Jacobsson, 2016a). In other words, these administrative tools offer lenses through which individuals are made “legible” and “measurable” for a state and therefore – as indicated by James Scott (1998) and Nicolas Rose (2005) in their seminal books – makes it possible to govern them from distance. Interestingly, legibility tools themselves, disguised as merely technical devices, often remain invisible (Bowker & Star, 2000) and escape public and academic attention.

In this paper we take a closer look at one of such legibility tools, a supposedly neutral and purely technical information technology for profiling the unemployed. More specifically we study a computerized scoring system which was used by frontline workers of Public Employment Offices (pl. powiatowe urzędy pracy) to profile all of the unemployed persons in Poland. Profiling involved dividing the unemployed population into three categories after a computer-based interview. Individuals were turned “into ranked and rated objects” (Citron & Pasquale, 2014, p. 3) through a scoring of their presumed “employability” (pl. potencjał zatrudnieniowy). The measurement of employability significantly affected life situations and professional chances of the unemployed citizens because depending on a categorization, different active labor market policies (ALMP) were accessed and different case management procedures applied[2].

We argue that the profiling technology served to shape the conduct (Rose 1999: 52) of the unemployed and imposed upon them an “ideal type of what a ‘normal’ citizen should be” (Wedel, Shore, Feldman, & Lathrop, 2005, p. 37). We propose to treat the profiling tool as a source of information about what was expected from the unemployed citizens by state authorities rather than verifying if it was (or wasn’t) capable of actually measuring “employability.” By analyzing a profiling questionnaire and scoring mechanism, we reconstruct what behaviors and attitudes were perceived by the state as “appropriate” and “desirable,” and which, on the contrary, were regarded as “wrong” and “demanding adjustment.” We reconstruct the implicit way of thinking and normative assumptions behind the criteria that were used to distinguish between the unemployed persons and to sort them into newly created target groups called “assistance profiles.” In this regard, this paper contributes to the strand of literature that departs from the functional analysis of public policy and places policy instruments in the center of empirical inquiry (Lascoumes & Le Gales, 2007). Those instruments are not merely technical and purely neutral devices, but rather they are bearers of values, fuelled by specific interpretations of citizenship which are used to “normalize” particular kinds of attitudes and behaviors (Lascoumes & Le Gales, 2007; Scott, 1998; Bowker & Star, 2000; Wedel, Shore, Feldman, & Lathrop, 2005, p. 37–38).

The article is structured as follows. First, we provide background information concerning the   implementation of profiling in Poland. Second, we present how we accessed data on profiling, what types of data we relied on, and how we analyzed them. Third, we reconstruct the normative assumptions that are hidden in the computerized profiling tool. In the conclusion, we explain how these assumptions are connected to the broader transformations of the welfare state.

2           Implementation of Computerized Profiling in Labor Market Policy in Poland

The standardized and computer-integrated classification system officially called “profiling the assistance for the unemployed” was obligatorily used in all Polish PES between 2014 and 2019. Profiling replaced the former system of allocation of ALMP which was based on broad target groups considered to be in particularly difficult circumstances. Beforehand, the vulnerability of the unemployed person was assessed by frontline workers of PES in reference to characteristics defined in a legal framework (such as long-term unemployment, young or old age, single parenthood). The previous system was criticized as non-transparent, leaving space for street-level bureaucrats’ discretion as well as leading to territorially differentiated policies and creaming (giving preferential access to ALMP to the higher skilled unemployed) (for the analysis of its actual working, see Sztandar‐Sztanderska, 2009; Sztandar-Sztanderska, 2016; Garsten et al., 2016).

The profiling system introduced in 2014 was legitimized through an expert and managerial narrative as a solution to those problems (Kwiatkowski, 2015; Męcina, 2015; Wiśniewski & Wojdyło-Preisner, 2015). The introduction of an algorithm was presented as a way to standardize assessments of the unemployed all over the country, unify the principles of allocation of ALMP, and make spending in this respect “rational” and “efficient” (for critical accounts, see: Sztandar-Sztanderska, 2013; Niklas, Sztandar-Sztanderska, & Szymielewicz, 2015; Sztandar-Sztanderska & Zieleńska, 2018). As a result of profiling, all the unemployed were ranked into three categories (“assistance profiles”) depending on the calculated value of their presumed employability. In order to make this calculation, software used client data filled in during a registration process and interview, conducted by a client counsellor with the help of an electronic template. However, the questionnaire and the scores, as well as the logic behind classification were kept secret by Ministry officials. They were not included in the law, nor discussed during the legislative process, and the Ministry of  Labor and Social Policy (from 2015 the Ministry of Family, Labor and Social Policy) refused several times to make them available to researchers, activists, and the general public.

The only thing that what was known from the legal act was the fact that the unemployed were divided into three groups, called first, second, and third assistance profile, and that categorization was consequential in terms of access to active labor market policies and case management procedures. The legal act did not specify who was included in what group – the criteria of the assessment and distribution into profiles were inscribed in the profiling tool. Later, from the leaked internal materials of Ministry of Labor and Social Policy, it was possible to learn that the first profile was supposedly regrouping people with the highest level of employability (Ministerstwo Pracy i Polityki Społecznej, 2014), the second profile was meant for those with average scores, and the third profile was for people with the lowest employability.

In terms of the ALMP, priority was given to people included in the second assistance profile who were allowed to apply for the largest range of services including, among other things, job placement and counselling, vocational training, apprenticeships, subsidies for opening their own business, and various forms of subsidized employment (Sejm, 2014, article 33 paragraph 2c). Unemployed people with a supposedly high level of employability included in the first assistance profile were allowed to use mainly the job placement services. Such a restriction was presented as a measure against allocating resources to those with the highest chances for taking up work. However, one has to be skeptical when it comes to its supposedly anti-creaming effect of this restriction since PES still had the legal possibility to offer them some of the services intended for the second profile such as training or subsidies for opening their own business provided that a particular case of a person could be considered “justified.”

In regards to the persons classified into the third profile, they were not entitled to any of the regular instruments offered by PES. Instead, according to the law, they could have been referred to outsourced services and two types of facultative programs. The problem was that these measures were hardly accessible. Outsourcing was very limited in scope, and facultative programs were rarely launched by PES (Herman-Pawłowska et al., 2016). Therefore, people profiled as lacking “employability” were, in practice, excluded from most forms of active labor market programs (Sztandar-Sztanderska, 2013; Niklas et al., 2015; Sztandar-Sztanderska & Zieleńska, 2018). Problems of access to ALMP might have affected approximately between 3 and 4 hundred thousand people a year, if we count all persons included in this group (own calculations based on: Ministerstwo Rodziny, Pracy i Polityki Społecznej, 2019).

The exclusion from ALMP was considered, perhaps, the most controversial effect of the profiling. The Supreme Audit Office found it “discriminating” and in contradiction to “the right to equal treatment” (Najwyższa Izba Kontroli, 2019) which was one among many reasons why the profiling was withdrawn in 2019. It should be noted, however, that being in the third profile held one important advantage from the perspective of the unemployed: access to free healthcare insurance with the exemption of job search requirements and sanctioning[3]. For people who were assigned the third profile, healthcare became an unconditional right, whereas for others, it remained conditional upon job search requirements and participation in ALMP.

3           The Study of Profiling the Unemployed

In this section we will describe how we accessed data on profiling, what types of data we relied on, and how we analyzed them.

3.1      Access to Data

In the case of profiling, similarly to other ICTs based on scoring (Citron & Pasquale, 2014; Pasquale, 2015; O’Neil, 2016), the classification criteria used to distinguish between the individuals, assess them, and classify them into groups were non-transparent and difficult to obtain. As previously mentioned, legal acts regulating profiling provided only basic information – largely that the unemployed would be divided into three assistance profiles and that people included in each category were, on the one hand, potentially entitled to a different range of services and support and had different obligations, on the other (Sejm, 2014). The crucial bits of information, including the questionnaire that was used to gather information, criteria of differentiation, the rationale behind it, and the scoring mechanism, were, however, guarded as a secret by the Ministry of Labor and Social Policy. They were concealed even though they actually defined who can get the services specified by the law which is a political decision par excellence (Lasswell, 1950).

This politically sensitive information has been publicly disclosed thanks to an electronic journal Dziennik Internautów and the involvement of – a non-governmental activist organization – Panoptykon Foundation. Dziennik Internautów (Maj, 2014) first published leaked electronic questionnaire and internal ministerial materials intended exclusively for PES staff (Ministerstwo Pracy i Polityki Społecznej, 2014). Then, the Panoptykon Foundation, with whom one of us conducted pilot research on profiling (Niklas et al., 2015), decided to challenge the Ministry of Labor and Social Policy before the court after their refusal to reveal the algorithm. Lawyers from Panoptykon Foundation sued the Ministry of Labor and Social Policy for refusal of granting access to the algorithm as public information and won this case in court. It enabled us to break the black box of the computerized scoring system which we analyzed, for the purpose of this article, from the perspective of values and ideas embedded in it.

3.2         Research Questions

Our aim was to shed light on the normativity of a supposedly technical and neutral profiling instrument, by asking whether and how values, convictions, and prejudices of the designers of the profiling tool were – as Cathy O’Neil phrased it – “camouflaged” within the technology (O’Neil, 2016), and, more specifically, what kind of ideal of citizen the tool imposed on the unemployed (Wedel et al., 2005). Our research questions were: what behaviors, attitudes, and circumstances were perceived – according to the profiling tool – as “appropriate” and “desirable” for the state, and which, on the contrary, were regarded as “wrong” and “demanding adjustment”? What were the implicit ways of thinking and normative assumptions behind the criteria that were used to distinguish between the unemployed persons and sort them into newly created target groups called assistance profiles?

3.3      Types of Data and Methods of Analysis

Findings presented in this article are mostly based on the analysis of the profiling classification instrument itself: i.e. a profiling questionnaire and the scoring mechanism. However, in order to understand how both the questionnaire and scoring mechanism were developed and what role they played in labor market policy, we also reconstructed the policy-making process. For this purpose, first, we analyzed the legal framework that regulated profiling (Sejm, 2014; Minister Pracy i Polityki Społecznej, 2014) and the legislative processes that lead to it. Second, we conducted 4 in-depth interviews with the people involved in creating the profiling tool, and analyzed the available statistics, policy making documents (including ministerial manual for PES (Ministerstwo Pracy i Polityki Społecznej, 2014), statistical justification of decisions behind the algorithm (Młodożeniec, 2013, 2014d, 2014c, 2014b, 2014a), and the official answers to our questions provided by the Ministry of Labor and Social Policy who agreed to reply after the court ruling. However, this data served only as background information, as this particular paper focuses on the normative ideal of citizen prescribed by the profiling tool.

The main type of data analyzed herein was the profiling instrument composed of the computer-integrated standardized questionnaire and scoring system on the basis of which the unemployed were distributed into groups[4]. Both the questionnaire and the scoring turned out to be simplistic, having little, if nothing, in common with mathematical models that apply data mining techniques and/or involve machine learning in order to make predictions about the future (for an analysis of use of such a technique in social policy, see Dubois, Paris, & Weill, 2018). Instead of relying on immense data sets, data used in Poland for profiling the unemployed were quite modest. The data set was restricted to standardized responses to a questionnaire composed of 24 questions. Eight questions relied on data gathered during registration with PES, such as age, sex, education level and work experience, foreign language skills, duration of unemployment, disability, and registered instances of non-compliance with PES requirements (meaning refusals of suitable employment or no-shows for an appointed meeting). The rest of the pre-defined answer choices had to be filled into the software by the client counsellors who were responsible for conducting the interviews with the unemployed.

The questionnaire used to profile the unemployed was developed mostly by a group of civil servants working in PES and the Ministry of Labor and Social Policy, and, only later, slightly revised by a statistician. Their convictions and ideas about how an unemployed person should and should not behave, what he or she can or cannot reasonably expect from the state, and what characteristics make him or her more employable are visibly reflected in the tool. The first step of analysis was to thematically code all the questions and answers from the profiling questionnaire to identify what circumstances, experiences, behaviors, attitudes, and characteristics were perceived as important from the state’s perspective. This thematic coding helped us to reconstruct what constitutes the ideal unemployed person imposed by the tool. We have distinguished three dimensions around which the normative assumptions underlying the profiling tool are grouped: (i) the division of responsibility between the state and the citizen, (ii) motivation to work, and (iii) key structural or socio-demographic characteristics.

Second, we analyzed the algorithm which turned out to be a simple mechanism that assigned scores to each of standardized answers, summarized them, and depending on the total, sorted the unemployed person into one of the three assistance profiles[5]. Every response had an assigned score, ranging from 0 to 8. Receiving 0 means the highest employability possible while any result beyond zero might be interpreted as a deviation from the ideal of employability for the unemployed. The higher the score the more a person deviates from this ideal. People receiving in total more than 22 points were grouped into the second profile with a supposedly average level of employability. If the total score was higher than 59, a person was assigned to the third profile for people with low levels of employability. The maximum score from the overall questionnaire was 155 points. The minimum was 0, however, this score was only possible for men to achieve because 1 point was automatically tallied to all women respondents.

4           What Makes an Ideal Unemployed Person?

4.1      The Division of Responsibility Between the State and the Citizen

We have distinguished three dimensions that make up an ideal unemployed person. The first refers to the division of responsibility between the state and the citizen, as far as solving the problem of unemployment is concerned. This aspect relates to convictions and ideas embedded in the profiling tool about what an unemployed person can or cannot reasonably expect from the state, and what the desired scope of responsibility of an individual and of the state is when it comes to dealing with unemployment.

Table 1: Answers indicating illegitimate or legitimate reasons for registering in PES

Question

Answers

Scores

15. What is the main reason that you have registered with the employment agency?

obtaining a preretirement benefit

8

applying for support at the social assistance center

5

obtaining health insurance

4

securing an entitlement to unemployment benefits

2

help getting work in the sense of receiving offers of suitable employment

0

help from the employment agency such as apprenticeship, training, etc.

0

23. Would you register at the employment agency while already having an entitlement to/source of health insurance?

No, I wouldn’t

4

Yes, I would

0

 

There were a number of questions in the profiling questionnaire that were used to distinguish legitimate and illegitimate expectations of the state. Decision to turn to the PES for support was considered wrong when a person wanted financial support or health insurance rather than job placement and ALMP (question no 15). All responses indicating such expectations were penalized by points, although the exact scoring depended on a specific type of financial support a person registered for (see Table 1). Preretirement benefits were considered the worst for the employability of an individual (8 points). People declaring the need for social assistance benefits were given slightly less points (5 points). Those who wanted an unemployment benefit deviated even less from the ideal of the employable individual (2 points), yet this expectation was still decreasing their employability in the eyes of the state. People who declared that their main reason for registration was for health insurance were penalized by 4 points; additionally, all the unemployed risked receiving another 4 points if they also declared that – in having another option of health insurance (e.g. being insured by a spouse or parents in the case of the youth) – they would not register with PES at all. By contrast, willingness to participate in ALMP or to use job placement services were considered legitimate reasons for registration with PES. Those answers were scored “0.”

The profiling questionnaire also implied what an unemployed person could or could not reasonably expect from the PES in terms of assistance in finding a job (table 2). A person who thought that it was the task of PES to find him or her a job was seen as departing from accepted standards. Such an answer scored 7 points, and in contrast, a belief in the possibility of finding a job, either with the help of PES or on their own, was treated as a desirable attitude.

Table 2: Answers indicating appropriate or inappropriate expectations towards PES

Question

Answers

Scores

14. Do you expect to find work on your own in the near future?

I have no chance of finding work – the employment agency should find me work

7

I have a chance of finding work, but I need support from the employment agency

0

I have a good chance of finding work on my own

0

 

To summarize, the profiling tool implies a very specific understanding of the relationship between the state and the unemployed persons. Individuals shouldn’t expect to have a right to so-called “passive labor market policies,” i.e. any type of benefits or healthcare. It means that the citizens are required to take considerable responsibility for finding a job and for supporting themselves financially. The profiling questionnaire makes them aware of the limited responsibilities of the state in response to unemployment.

4.2      Motivation to Work

Another dimension important for the reconstruction of the ideal underlying the profiling tool was the motivation to work. Profiling was supposed to (i) determine whether someone is motivated enough to take up work, and (ii) distinguish people with the appropriate types of motivation from others with inappropriate ones.

In the tool, there were several questions and answers designed to test whether an unemployed person showed motivation to find a job. Question 15 about reasons for registering with PES was already mentioned in the previous section. As shown in Table 1, a person who registered in order to make use of job placement or ALMP was seen as behaving in an appropriate way as opposed to people seeking financial support.

There were also other questions directed towards the assessment of motivation – they related to several issues: whether a person is at all motivated to take up work (question no 21 and 13) and is looking for it (13), how soon a person is ready to start a job (22), what actions has she or he undertaken in order to find it (17), and what is she or he able to do in order to find it (18). Table 3 shows all the answers that were treated as a sign of lack of motivation. Most of them were worth 8 points which is the maximum a person could get for a single answer. It means that their significance for the overall measurement was high and they were treated as deviations from the ideal. A scoring model negatively assessed people who were not able to tell when they could start work; did not send a CV or take other actions to find a job; and were reluctant to intensify job search activities or to do other things to increase chances of employment.

 Table 3: Answers indicating lack of motivation to work

Question

Answers

Scores

18 What are you able to do in order to improve the chances of your finding work?

 

I am not prepared to do anything (among others: not ready to be more active in looking for work)

8

21. What inclines you to take up work, apart from the income?

Nothing inclines me to take up work (not even the income)

8

13. Please indicate the reasons why it is difficult for you to find work.

Not convinced of the need to take work / not looking for work

8

22. When are you available to start work?

I’m not able to tell

8

17. In the last month have you independently prepared material for a job application (CV, cover letter)?

No

4

No, because it has not been necessary (3 points)

3

 

A number of answers in the questionnaire were also directed to single out people whose answers indicate motivation to take up work. As shown in Table 4, apart from the already mentioned appropriate reasons for registering (15), the properly motivated unemployed should be looking for work on their own (16), should be able to point to at least three job search activities from a long list of options (16), and declare the readiness to take up work immediately (22).

Table 4: Answers indicating motivation to work

Question

Answers

Scores

15. What is the main reason that you have registered with the employment agency?

Help getting work in the sense of receiving offers of suitable employment

0

16 Are you looking – or have you been looking – for work on your own?

Yes

0

16 What are you doing – or have you been doing – in order to find work?

At least 3 of the following options are required:

a) I send out CVs with a cover letter

b) I look at job adverts in the newspapers

c) I look at job adverts online

d) I call employers

e) I arrange to meet employers

f) I get help from the employment agency

g) I get help from a recruitment agency

h) I get help from friends and family

i) I attend job fairs

j) I try and find work experience or an internship, even if it is unpaid

k) I do voluntary work

0

17. In the last month have you independently prepared material for a job application (CV, cover letter)?

Yes

0

22. When are you available to start work?

Immediately

0

 

Moreover, the profiling tool distinguished between different sorts of motivation to work, implicitly assuming that some of them better than others. The details on this subject are presented in Table 6. Interestingly, according to the designers of the profiling tool, income should not be the only reason for looking for employment. Answering that money is the only factor inclining the unemployed to find a job was penalized with an added 2 points. What’s more, one should not be picky in regards to the salary – not agreeing to take up work for minimum wage, or agreeing to it only under certain conditions were seen as deviations from the ideal and scored with 7 points and 1 point respectively. Moreover, showing any external motivation to find work was also treated as opposed to the ideal, whereas the need to support family was scored with 1 point, admitting that expectations of other people were an essential motivator was scored with 6 points. Similarly, willingness to take up employment in order to gain retirement rights also did not match the ideal and was penalized with 2 points. As far as the appropriate motivation is concerned, the ideal citizen should be internally driven by the urge to be active and by the idea of professional and personal development.

Table 5: Answers indicating appropriate or inappropriate motivation

Question

Answers

Scores

21 What inclines you to take up work, apart from the income?

Keeping active

0

The prospect of professional and personal development

0

The necessity of providing for myself / my family

1

Securing a pension

2

Nothing, apart from the income, inclines me to work

2

Other people expect me to work

6

19.       Given the choice between taking work at the minimum wage and remaining out of work, what would you choose?

To take the work

0

To take the work under certain conditions (depending on the location, the type of work, whether it’s in your profession, what the possibilities are for professional development etc.)

1

To remain out of work

7

 

To summarize, ts shows that motivation to work is a key element of the ideal unemployed person in a twofold sense. On the one hand, the unemployed person is required to prove his or her motivation to work. Not showing motivation is considered as departing from the accepted standards. On the other hand, not all types of motivation are equally valuable. The appropriate motivation should have internal rather than external sources – instead of satisfying financial needs, a person should act upon his or her need of self-development and self-improvement. Interestingly: answering to the expectations and needs of other people (e.g. willingness to support family or satisfy others) is implicitly treated as an inappropriate incentive. What is quite striking here is that the ideal encapsulated in the profiling tool includes a very specific concept of an individual as someone who is socially disembedded and self-reliant.

4.3     Socio-Demographic Characteristics

The questionnaire also includes a number of questions connected to socio-demographic characteristics. They relate to gender, disability, place of residence, age, education, and duration of unemployment. Table 6 displays extreme characteristics in each category, i.e. those considered to be matching the ideal of employability and those the most distant from it. Higher education (0 points) is contrasted with having lower-secondary education or below (8 points); being unemployed for less than 6 months (0 points) is contrasted with being unemployed for more than 24 months (6 points), being between 25 and 39 years old (0 points) is contrasted with being older than 49 years old (5 points), having no disability (0 points) is contrasted with having severe disability (5 points), living in urban agglomeration or a large city with suburbs is contrasted with living in remote rural areas (5 points), and finally, being a man (0 points) is contrasted with being a woman (1 point).

Table 6: Socio-demographic characteristics with minimum and maximum scoring

Question

Answers

Scores

3. Education of the unemployed person

 

Up to lower-secondary education

8

Higher

0

7. Length of time the unemployed person has been out of work

24 months or longer

6

Less than 6 months

0

1. Age of the unemployed person

50 or over

5

30 – 39

0

25 – 29

0

6. Disabilities of the unemployed person

An individual with a serious disability

5

Healthy individual (no medical certification)

0

11. Place of residence in terms of distance from potential places of employment.

A village or settlement located at a considerable distance from the labor market

5

A large town or city with suburbs

0

Urban area

0

2.         Gender of the unemployed person

female

1

male

0

 

The scoring suggests that the person who embodies the ideal is a well-educated male between 25-39, from a large city, in good health, who remained jobless for less than 6 months. What is even more striking, however, is the fact that those characteristics are derivatives of certain structural problems such as unequal access to education, discrimination of women in the labor market and their relatively low employment levels, residual public transportation in rural areas, scarce support systems for disabled people both in terms of social and vocational integration, and simply regional economic inequalities. Yet, those structural characteristics are converted by the profiling tool into features of individuals that influence their assessment as an unemployed person.

5           Conclusions

In this article, we proposed to look at the ICT for profiling the unemployed implemented in Poland in between 2014 and 2019 from a specific perspective of a legibility tool. Our analysis focused on reconstructing the set of normative criteria underlying this technology, against which the unemployed people had been assessed. These criteria are related to a certain concept of social citizenship according to which, access to benefits and services should no longer be based on collective statuses or payments of social contributions, but on adequate behaviors and attitudes of the individual. In the analysis, we determined three dimensions which form the ideal of social citizenship encapsulated in the profiling tool. Each of the highlighted dimensions relates directly to the key aspects of a welfare state transformation, namely, to the new social contract, the new concept of a citizen, and the new perception of social risks (e. g. Clarke, 2005; Garsten & Jacobsson, 2004, 2016b; Serrano Pascual & Magnusson, 2007).

As shown in our analysis of the questionnaire and the scoring system, the division of responsibility between the state and the citizen is founded on the assumption that the citizens may legitimately count on and ask for access to ALMP or support in finding a job, but not on the state finding a job for them. It means that an active engagement in the process of job search and adjusting skills to labor market demands is expected from the citizen; whereas, counting on any kind of benefits/insurance from the state is considered an illegitimate claim (Portet & Sztandar-Sztanderska, 2010). This conceptualization of the share of tasks and duties relates directly to what Serrano-Pascual and Magnusson (2007) call the new social contract, marking a shift from a paradigm based on provision of welfare to a paradigm based on provision of activation. As a result, the role of the state is no longer to protect the workforce from the negative effects of market forces by granting them benefits and insurance, but to help them adapt to market expectations through activation policies. In this way, the state becomes responsible for the “moral-therapeutic regulation of behaviour” (Serrano Pascual & Magnusson, 2007) meaning that it is about correcting individual behaviors and attitudes rather than providing a safety net to minimize risks generated by the market. It is the “self” of the citizen which becomes the constant object of the state’s interest (Rose, 1999; Garsten & Jacobsson, 2016a). This paradigm shift also produces a specific understanding of the citizen as well as new perception of social risks which correspond to other dimensions identified in the analysis of the profiling tool.

The second dimension identified in the analysis, relates to the motivation to work. On the one hand, the “ideal” or “normal” unemployed citizen should be showing motivation to find a job and be ready to take active part in the process. On the other hand, there is a strong expectation for this motivation to be internally driven and founded on the need for self-development and the urge to be active. Such normative assumptions are also at the core of the concept of the citizen characteristic of the paradigm based on provision of activation, i.e. the employable individual (Garsten & Jacobsson, 2004, 2016b; Serrano Pascual & Magnusson, 2007). The success of the new social contract is strongly dependent on the self-reliance of the individual and whether she or he takes the responsibility for working on one’s own attractiveness for the labor market. It means that the individual is considered to be both – the source of problems (because of wrong attitudes and skills) and the source of solutions (understood as becoming entrepreneurial, and working on oneself in order to improve and gain independence).

Finally, the reconstructed ideal also includes certain socio-demographic characteristics which are desired – they relate to gender, disability, place of residence, age, education, and duration of unemployment. This means that the profiling tool converts certain structural problems into features of individuals that undergo assessment (e.g. a person’s low education is a problem not educational inequalities; a person’s gender is a problem, not the fact that women are discriminated against in the labor market). In this way, social risk becomes individualized by shifting the diagnosis of the causes of certain problems from social and economic inequalities or the underperformance of the economy to the maladjustment of the individual.

 

References:

Berkel, R. van, & Valkenburg, B. (Eds.). (2007). Making it personal: Individualising activation services in the EU. Bristol, UK: The Policy Press.

Betzelt, S., & Bothfeld, S. (Eds.). (2011). Activation and labour market reforms in Europe: Challenges to social citizenship. Houndmills, Basingstoke, Hampshire; New York: Palgrave Macmillan.

Bowker, G. C., & Star, S. L. (2000). Sorting Things Out: Classification and Its Consequences. Cambridge Mass.: MIT Press.

Caswell, D., Marston, G., & Larsen, J. E. (2010). Unemployed citizen or ‘at risk’ client? Classification systems and employment services in Denmark and Australia. Critical Social Policy, 30(3), 384–404.

Citron, D., & Pasquale, F. A. (2014). The Scored Society: Due Process for Automated Predictions. Washington Law Review, 89(2014–8).

Clarke, J. (2005). New Labour’s citizens: Activated, empowered, responsibilized, abandoned? Critical Social Policy, 25(4), 447–463.

Dubois, V. (2009). Le paradoxe du contrôleur: Incertitude et contrainte institutionnelle dans le contrôle des assistés sociaux. Actes de la recherche en sciences sociales, 178(3), 28.

Dubois, V. (2014) The State, Legal Rigor, and the Poor: The Daily Practice of Welfare Control, Social Analysis, 58(3), 38–55.

Dubois, V., Paris, M., & Weill, P.-E. (2018). Targeting by Numbers. The Uses of Statistics for Monitoring French Welfare Benefit Recipients. In L. Barrault-Stella & P.-E. Weill (Eds.), Creating Target Publics for Welfare Policies (pp. 93–109). Cham: Springer International Publishing.

Garsten, C., & Jacobsson, K. (Eds.). (2004). Learning to be employable: New agendas on work, responsibility, and learning in a globalizing world. Basingstoke: Palgrave Macmillan.

Garsten, C., & Jacobsson, K. (2016a). Sorting people in and out: The plasticity of the categories of employability, work capacity and disability as technologies of government. ephemera. theory & politics in organization, 13(4), 825–850.

Garsten, C., & Jacobsson, K. (2016b). Transparency as ideal and practice. Labour market policy and audit culture in the Swedish public employment service. Statsvetenskaplig tidskrift, 118(1), 69–92.

Garsten, C., Jacobsson, K., & Sztandar-Sztanderska, K. (2016). Negotiating social citizenship at the street-level: Local activation policies and individualization in Sweden and Poland. In M. Heidenreich & D. Rice (Eds), Integrating Social and Employment Policies in Europe (pp. 265–294), Cheltenham, UK: Edward Elgar Publishing.

Góra, M. (2006) Active labor market policies in Poland. In J. Kluve (ed.), Active labor market policies in Europe: Performance and perspective, (pp. 122-133), Berlin: Springer.

Herman-Pawłowska, K., Stronkowski, P., Bienias, S., Dybaś, M., Kolczyński, M., Kulawik-Dutkowska, J., & Skórska, P. (2016). Ocena skutków regulacji wybranych aspektów wdrażania Ustawy z dnia 14 marca 2014 r. O zmianie ustawy o promocji zatrudnienia i instytucjach rynku pracy oraz niektórych innych ustaw. Retrieved November 2019 from http://ideaorg.eu/file.php?i=podstrony/50a4f51c7d20eb54f2e932940448c6fc.pdf

Kwiatkowski, E. (2015). Podstawy teoretyczne zmian w polskiej polityce rynku pracy. Polityka Społeczna, numer tematyczny 2, 13–16.

Lascoumes, P., & Le Gales, P. (2007). Introduction: Understanding Public Policy through Its Instruments? From the Nature of Instruments to the Sociology of Public Policy Instrumentation. Governance, 20(1), 1–21.

Lasswell, H. D. (1950). Politics: Who gets what, when, how. New York: P. Smith.

Maj, M. (2014). Profilowanie bezrobotnych—Ujawniamy dokument z instrukcjami dla urzędników (#tylkounas). Dziennik Internautów. Retrieved November 2019 from   http://di.com.pl/profilowanie-bezrobotnych-ujawniamy-dokument-z-instrukcjami-dla-urzednikow-tylkounas-50960

Męcina, J. (2015). Reforma instytucjonalna rynku pracy—Jej pierwsze efekty i kolejne wyzwania. Polityka Społeczna, numer tematyczny 2, 2–12.

Minister Pracy i Polityki Społecznej. Rozporządzenie Ministra Pracy i Polityki Społecznej z dnia 14 maja 2014 r. W sprawie profilowania pomocy dla bezrobotnego., Pub. L. No. Dz.U. 2014 poz. 631 (2014).

Ministerstwo Pracy i Polityki Społecznej (2014). Profilowanie pomocy dla osób bezrobotnych. Podręcznik dla pracowników powiatowych urzędów pracy.

Ministerstwo Rodziny, Pracy i Polityki Społecznej (2019). Statystyki strukturalne czerwiec 2019. Retrieved November 2019 from https://psz.praca.gov.pl/documents/10828/9119519/Statystyki%20strukturalne%20czerwiec%202019.zip/29658fcb-b481-46e4-9043-e586b4d5362d?t=1564131659000

Młodożeniec, M. (2013). Profilowanie pomocy dla osób bezrobotnych raport z analizy danych z pilotażu (Nr 1). Warszawa.

Młodożeniec, M. (2014a). Powody zmiany profilu pomocy raport z analizy danych jakościowych z profilowania osób bezrobotnych w 2014 roku (Nr 4). Warszawa.

Młodożeniec, M. (2014b). Profilowanie pomocy dla osób bezrobotnych raport końcowy z wyników profilowania  pomocy dla osób bezrobotnych w 2014 roku (Nr 5). Warszawa.

Młodożeniec, M. (2014c). Profilowanie pomocy dla osób bezrobotnych raport nt. Wyników profilowania pomocy dla osób bezrobotnych  analiza danych cząstkowych z 319 Urzędów Pracy (Nr 3). Warszawa.

Młodożeniec, M. (2014d). Profilowanie pomocy dla osób bezrobotnych raport z analizy danych z profilowania w 5 Urzędach Pracy: W Warszawie, Gdańsku, Jarosławiu, Nysie i Oświęcimiu (Nr 2). Warszawa.

Najwyższa Izba Kontroli (2019). Aktywizacja bezrobotnych—Ryba czy wędka? Retrieved November 2019 from https://www.nik.gov.pl/aktualnosci/aktywizacja-bezrobotnych-ryba-czy-wedka.html

Niklas, J., Sztandar-Sztanderska, K., & Szymielewicz, K. (2015). Profiling the Unemployed in Poland: Social and Political Implications of Algorithmic Decison Making. Retrieved November 2019 from https://panoptykon.org/biblio/profiling-unemployed-poland-social-and-political-implications-algorithmic-decison-making

O’Neil, C. (2016). Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. London: Penguin Book.

Portet S., Sztandar-Sztanderska, K. (2010) Poland. Unemployment benefit: haunted by a lack of legitimacy. In F. Lefresne (Ed.), Unemployment Benefit Systems in Europe and North America: Reforms and Crisis (pp. 261–280).

Pasquale, F. (2015). The black box society: The secret algorithms that control money and information. Cambridge: Harvard University Press.

Rose, N. S. (1999). Powers of freedom. Cambridge: Cambridge University Press.

Rose, N. S. (2005). Governing the soul: The shaping of the private self (2. ed., [reprint]). London: Free Association Books.

Scott, J. C. (1998). Seeing like a state: How certain schemes to improve the human condition have failed. New Haven: Yale University Press.

Sejm. Ustawa z dnia 14 marca 2014 r. O zmianie ustawy o promocji zatrudnienia i instytucjach rynku pracy oraz niektórych innych ustaw. Dz.U. 2014 poz. 598 (2014).

Serrano Pascual, A., & Magnusson, L. (Eds.). (2007). Reshaping welfare states and activation regimes in Europe. Brussels; Oxford: P.I.E.-Peter Lang.

Sztandar‐Sztanderska, K. (2009). Activation of the unemployed in Poland: From policy design to policy implementation. International Journal of Sociology and Social Policy, 29(11/12), 624–636. https://doi.org/10.1108/01443330910999069

Sztandar-Sztanderska, K. (2013). Nie zrzucajmy całej winy na nieefektywne urzędy pracy i fikcyjnych bezrobotnych. Uwagi o niepożądanych konsekwencjach nowelizacji Ustawy o promocji zatrudnienia i instytucjach rynku pracy. Retrieved November 2019 from http://www.wrzos.org.pl/download/Dodatkowa%20ekspertyza%20prawna.pdf

Sztandar-Sztanderska, K. (2016). Obywatel spotyka państwo: O urzędach pracy jako biurokracji pierwszego kontaktu. Warszawa: Wydawnictwo Naukowe Scholar.

Sztandar-Sztanderska, K., & Zieleńska, M. (2018). Changing social citizenship through information technology. Social Work & Society, an International Online Journal, 16(2).

Wedel, J. R., Shore, C., Feldman, G., & Lathrop, S. (2005). Toward an Anthropology of Public Policy. The ANNALS of the American Academy of Political and Social Science, 600(1), 30–51.

Wiśniewski, Z., & Wojdyło-Preisner, M. (2015). Profilowanie bezrobotnych w Polsce i Niemczech. Polityka Społeczna, 22–27.

Authors Addresses:

Karolina Sztandar-Sztanderska, Ph.D.
University of Warsaw, Poland
Institute of Sociology
Karowa 18, 00-927 Warsaw
k.sztanderska@gmail.com

Authors Addresses
Marianna Zielenska, PhD.
University of Warsaw, Poland
Institute of Sociology
Karowa 18, 00-927 Warsaw
marianna.zielenska@gmail.com 



[1] The authors would like to particularly thank Christina Garsten, Kerstin Jacobsson, Kristina Tamm Hallström and Martin Heidenreich, whose previous work, insightful comments as well as long-term academic support inspired this line of research. We would also like to thank Panoptykon Foundation for fighting the access to algorithm in court, in particular Jedrzej Niklas (currently employed at Data Justice Lab, Cardiff University), Wojciech Klicki and Katarzyna Szymielewicz. The article was prepared in the frame of the ongoing project “Information technologies in public policy. Critical analysis of the profiling the unemployed in Poland” directed by Karolina Sztandar-Sztanderska and financed by the National Science Centre, Poland (2016/23/B/HS5/00889). We would like to thank other research team members: Alicja Palęcka, Michał Kotnarowski and Barbara Godlewska-Bujok. For the purpose of the paper, we also used data gathered in the frame of previous small pilot projects and one scholarship: 1) “To measure, to weigh, to calculate, to classify. Methods of making citizens legible” carried out by K. Sztandar-Sztanderska (DSM 112900/16); 2) “Profiling the unemployed - social and political consequences of new categorization tools introduced by local labour offices” carried out by K. Sztandar-Sztanderska (DSM 110400/66) and M. Zieleńska (DSM 110 400/72); 3) “Profiling the Unemployed in Poland. Social and Political Implications of Algorithmic Decision Making” research conducted by K. Sztandar-Sztanderska and J. Niklas for Panoptykon Foundation and funded by Media Democracy Fund, Ford Foundation and Open Society Foundations.; 4) Scholarship for outstanding young scholars granted to K. Sztandar-Sztanderska by Polish Ministry of Science and Higher Education for years 2015-2018 (653/STYP/10/2015).

[2] Frontline workers were allowed to change the automatic classification, but – according to data from the first six months of profiling – this possibility was only exceptionally used (Sztandar-Sztanderska & Zieleńska, 2018, p. 4).

[3] In contrast to other countries, registration with PES in Poland gives people access to free healthcare insurance rather than to unemployment benefits (Góra 2006; Portet & Sztandar-Sztanderska, 2010). Benefits have been accessible for less than 17 per cent of the unemployed (data for 2010-2019 from: Ministerstwo Rodziny, Pracy i Polityki Społecznej, 2019).

[4] Frontline worker could have changed the assistance profile, if she or he disagreed with the automatic classification of the unemployed person. However – according to data from the first years of profiling, this possibility was used only in 0,58% of cases of profiling (Młodożeniec, 2014a, s. 6; Sztandar-Sztanderska & Zieleńska, 2018, s. 4).

[5] At the early stage of preparing a profiling tool, there was no numerical scoring at all. Instead, each answer was assigned a symbol representing one of three assistance profiles (a, b, c). Depending which answers „a“, „b“ or „c“ prevailed, an unemployed was supposed to be categorised respectively into first, second or third assistance profile. Only later the Ministry decided to implement scoring. The numerical values were assigned and then calibrated in a way that a value of Alfa Cronbach indicator will increase. We will however not go into details in this paper as to how the profiling tool was developed and what was problematic about the algorithm. These questions will be addressed in other publications.