1 Introduction

Teacher education in higher education plays a central role in educational policies aimed at improving the quality and equity of teaching practice in the classroom. Although academic qualifications alone are insufficient to evaluate teacher effectiveness, completion of an undergraduate degree program in the area they teach is associated with recognised factors for improving student educational outcomes, especially for the most disadvantaged students, such as mastery of content and effective adoption of teaching methods and resources (^{PAUL; BARBOSA, 2008}; ^{CARVALHO et al., 2021}). Therefore, it is not surprising that expanding specific higher education teacher education for basic education teachers is enshrined in goal 15 of the National Education Plan (^{BRASIL, 2014}), and recent regulatory initiatives for teacher education, such as the National Curriculum Guidelines for Initial Teacher Education for Basic Education and the Common National Base for Initial Teacher Education for Basic Education (BNC-Formação) (^{MEC, 2019}), aim to improve the quality of teacher education.

However, there is a gap between the completion of higher education teacher training courses and the actual exercise of teaching that affects not only the composition of the basic education teacher workforce, but also the outcomes that this group achieves in their professional insertion and trajectory. Firstly, like in other Latin American countries, the teaching profession suffers from structural deficiencies in Brazil, such as unsatisfactory remuneration and poor working conditions, which repel graduates who received quality initial training (^{VAILLANT, 2006}; ^{LOUZANO et al., 2010}). Secondly, being a predominantly female profession, professional insertion decisions are strongly influenced by family choices, such as aspects involving marriage and fertility (^{ALVES et al., 2016}). Lastly, the same reasons that reduce the attractiveness of the profession also lead many teachers to seek additional jobs or work in more than one school, which harms students' academic outcomes (^{ELACQUA; MAROTTA, 2019}).

The study of the professional insertion of graduates of teacher training courses can also shed light on aspects of the expansion of the higher education system in the country and its relation to inequalities in the job market for graduates. The intense transformations that have occurred both in the normative scope and in the evolution of the number of enrollments in higher education since the 1990s cannot be understood without observing the behaviour of the Education field, which has been deeply affected by legislative and normative changes concerning teacher training and career (^{HONORATO; VIEIRA; ZUCCARELLI, 2019}; ^{VAILLANT; MANSO, 2022}). The issues above become even more interesting when we consider that this same area of training has been relegated to the less prestigious spaces of the system (^{BROCK; SCHWARTZMAN, 2004}).

The present study draws on this literature to discuss to what extent teacher education in higher education is associated with different points of entry into the formal labour market and how the socioeconomic backgrounds of graduates condition this transition. Thus, we seek to take an additional step in relation to previous studies focused on the universe of graduates in higher education in Brazil (^{VIEIRA; HONORATO; RODRIGUES, 2022}; ^{VIEIRA, 2023}), by examining how specific qualitative differences in higher education career matter for the professional outcomes of graduates.

We assume that in order to better understand the effect of teachers' characteristics on the performance and progression of basic education students, it is essential to analyze the initial training of these teachers, which is increasingly taking place in higher education. To this end, we used data from the 2011 National Student Performance Exam (Enade), which for the first time in the evaluation's historical series allowed us to identify graduates from 14 different teacher training courses, as well as data on formal employment insertion from the 2014 Annual Social Information Report (RAIS).

Our analysis is organized into two stages. First, we examine the entry of graduates from Education courses into formal employment, with the aim of describing patterns of selectivity in the insertion into the occupational position focused on in this study. Next, we calculate the relationship between social origin and academic training characteristics and the chances of being employed in occupations considered typical or atypical for the Education field.

2 Previous Studies and Working Hypotheses

Although teacher education courses traditionally occupy subordinate positions in Brazilian higher education, the expansion of this stage in recent decades repositions the classic problem of literature on social stratification and equal opportunities through schooling. The traditional and patrimonial character of our higher education (^{PRATES; BARBOSA, 2015}) indicates difficulties in expanding this equality, which contributes to the distribution of students being more strongly associated with their social origin than their academic merit or occupational vocation. The effects of this limited openness are clearly visible in the perspective of young students, who greatly value their entry into higher education while realizing that it does not necessarily mean the fulfilment of their dream (^{BARBOSA; DWYER, 2016}).

Given the importance of teacher education programs for Brazilian higher education, both their institutional characterization, including which types of institutions, modalities, and shifts offer the courses, and the socioeconomic composition of enrollments, are fundamental elements to understanding to what extent the expansion of the system can result in some democratization. Enrollments in teacher education courses had the second highest growth in higher education between 1995 and 2015, which can be characterized by three processes (^{HONORATO; VIEIRA; ZUCCARELLI, 2018}): (1) the multiplication of enrollments in private institutions; (2) the increase in course offerings, especially in the North and Northeast regions; and (3) the exponential growth of distance education (DE) modality. If the current trend continues, the next generations of Brazilian teachers will be trained in the DE modality, mainly in private higher education institutions (^{SANTOS; LIMA; CARVALHAES, 2020}).

Similarly to what happens in other higher education systems (^{GERBER CHEUNG, 2008}), teacher training courses in Brazil traditionally cater to socially disadvantaged groups of student-workers, women, and those with low prior academic performance. Several analyses have shown that the Education area is among those where students from lower-income and less-educated families are more likely to graduate, suggesting that this audience is "redirected" by the system to courses of lower social value (^{BARBOSA; VIEIRA; TAGLIARI, 2017}; ^{KNOP; COLLARES, 2019}; ^{CARVALHAES; RIBEIRO, 2019}). Nevertheless, there are important qualitative differences among graduates of various teacher training courses, with Pedagogy graduates having a more disadvantaged family background than those of other teacher training courses, especially Mathematics and Natural Sciences (^{ZUCCARELLI; HONORATO; VIEIRA, 2018}).

In addition to the literature on the inequality of opportunities in access to and completion of undergraduate teacher education programs, our analysis is also based on two complementary research traditions: the first on mechanisms that explain the economic returns of higher education; and the second on teacher supply and the teaching job market. Both lines of study can be articulated by considering that the social self-selection for teacher education programs that we discussed in the previous paragraphs is directly related to the low monetary and social returns of graduates in the labour market. These studies, together with the characteristics of the teaching job market in Brazil, inform our hypotheses.

Studies aimed at understanding the mechanisms associated with graduates' outcomes in the labour market have indicated the importance of academic formation characteristics, such as the field of study and the skills acquired, and the graduates' professional trajectory, such as work experience during their studies, as explanatory factors for returns to undergraduate degrees (^{VIEIRA; HONORATO; RODRIGUES, 2022}). In line with the human capital theory, prior literature indicates that graduates' skills, especially those that are highly specific and related to well-defined technical profiles, explain a substantial portion of salaries in the labour market (e.g., ^{BOL; HEISIG, 2021}; ^{VAN DE WERFHORST, 2002}). Some studies show that particularly economic and technical skills related to practical work activities are associated with higher rewards compared to cultural skills (^{VAN DE WERFHORST; KRAAYKAMP, 2001}). Signalling theories, on the other hand, suggest that work experience can be rewarded in the market by indicating to employers, in addition to greater productivity, and lower training costs (^{AINA; CASALONE, 2020}). Regarding the effect of social background on salaries, previous studies show that family resources partially compensate for signalling failures in less occupationally specific areas of study, particularly at graduates' entry into the labour market (^{JACOB; KLEIN, 2019}).

The literature on the teaching job market has identified a series of factors, financial or not, associated with the individual choices of potential teachers, including salaries and benefits, working conditions, and school location (^{BETÉILLE; LOEB, 2009}). These factors affect the decisions of graduates from teacher education programs to choose teaching as a profession and, if they do, where they will seek to teach. Although studies on teachers in Brazil and Latin America point to characteristics that can increase the attractiveness of the teaching career, such as flexible work hours, vacation time, low unemployment rates, and a sense of altruism associated with the profession (^{OECD, 2006}; ^{VEGAS, 2005}), there are few financial incentives for the best high school students to choose teaching as a profession in Brazil (^{LOUZANO et al., 2010}). Reinforcing this scenario, research shows that working in typical occupations for newly graduated professionals in teacher education programs was not associated with higher incomes or higher quality courses (^{MACIENTE et al., 2015})¹.

National and international studies on teacher supply and career have indicated a relationship between teachers' academic background and a range of professional outcomes, including two that are of interest: (1) the choice between teaching and other careers and (2) the type of school for teaching. Students who choose or express a desire to teach tend to be less academically proficient (^{GOLDHABER; LIU, 2003}; ^{PODGURSKY; MONROE; WATSON, 2004}; ^{ALVES et al., 2016}) and, among those who choose teaching, those who enter the public school system through competitive exams are more likely to benefit from their families' cultural resources (^{CASTELLAR et al., 2010}). Graduates of more general courses, such as Pedagogy, seem to have a greater chance of working as teachers in public schools than their peers from more specific disciplines (^{BALLOU, 1996}), which may be reinforced by the greater supply of opportunities for these graduates in the public school system in Brazil. Additionally, there is evidence that uniform salary structures for the teaching profession lead to a shortage of teachers in specific areas, such as mathematics and science, as these individuals can access more lucrative job opportunities outside of teaching, including self-employment (^{MURNANE et al., 1991}; ^{VIEIRA, 2023}). Finally, several studies point to the regularity that more experienced and better-educated teacher candidates tend to choose schools with better working conditions and a socially advantaged student profile (^{GOLDHABER et al., 2014}; ^{BETÉILLE; LOEB, 2009}; ^{PAUL; BARBOSA, 2008}).

Based on the research traditions mentioned and the characterization of the teaching job market in Brazil, we outline some hypotheses about the outcomes of graduates in the teaching profession in the job market.

3 Methodology

Our data comes from the National Student Performance Exam (Enade) 2011 and the Annual Social Information Report (RAIS) 2014. Enade is an assessment of the performance of undergraduate students, who are divided into three large groups that take turns being evaluated each year. Thus, every three years, the same group of courses is reevaluated. In the exam, graduating students solve general training questions common to all areas and questions specific to their course. In 2011, bachelor's and teaching degree programs were evaluated separately for the first time by Enade. RAIS is an annual administrative record and mandatory declaration for all establishments in the country, containing information on the universe of formal employees in the country.

For this analysis, we obtained access to the restricted versions of the respective databases, which allow individuals to be identified by a unique key. We kept in the data the graduates who met two conditions: (1) were formally employed for two years or less; (2) did not complete postgraduate courses between the year of graduation and the year in which employment is measured. Both filters were performed in order to make the sample more homogeneous, eliminating possible confounding factors on the relationship between higher education and formal employment.

On the other hand, the data used has known limitations (^{OECD, 2018}). Among the problems with Enade, those related to the scope and consistency of information about enrollees deserve mention, which prevents a precise measure of the coverage of the evaluation in relation to the universe of higher education graduates. In addition, the evaluation suffers from problems in the development, selection, and use of test items that imply difficulties in the substantive interpretation of student results on Enade (^{SCHWARTZMAN, 2005}; ^{OECD, 2018}).

As a formal labour market record, the RAIS does not cover workers who are in the informal sector, self-employed or employers, which prevents its results from being interpreted as valid for the universe of Brazilian workers. Being an administrative record with information provided by employers, it should be interpreted with caution due to errors and omissions common to this format of information declaration (^{DE NEGRI et al., 2001}).

The frequency of missing information, especially in variables from the student socioeconomic questionnaire (such as gender, parents' education and race), is a known challenge for using Enade data (^{VIEIRA, 2023}; ^{RODRIGUES, 2024}). In addition to these fields, we also observe missing values in the variable that indicates the graduate's employment status and the score in the specific component of the exam. Table 2 describes the pattern of missing values by field of study. The variables with missing data present percentages of missing values between 10% and 25%. In general, the highest percentages are observed for courses in the areas of Humanities and Social Sciences and Languages. To avoid excluding cases with missing information, which could affect the representativeness of the sample, we followed the strategy adopted in other studies (^{VIEIRA, 2023}; ^{RODRIGUES, 2024}), opting to impute the values of missing observations through the Multivariate Imputation by Chained Equations method (RUBIN, 1996; ^{BUUREN; GROOTHUIS-OUDSHOORN, 2011}).

4 Results

5 Final Considerations

An important agenda for discussion on educational policies in Brazil emphasizes problems with the composition of the teaching workforce in the country, including the need to attract the best students to the teaching profession and to address deficiencies in the qualifications of current public school teachers. As a way of contributing to this debate, a significant set of studies has analyzed, on the one hand, who chooses the teaching profession, and on the other hand, how this professional group is characterized in the world of work. However, little is known about the "middle of the road": who are the graduates of teacher education programs who are actually entering the formal labor market in Brazil - in typical occupations such as teaching, or not.

In this article, we sought to at least partially fill this gap, contributing to an understanding of how the academic trajectory and social origin of graduates from teacher education programs are associated with their insertion into formal employment, particularly in teaching positions. In general, we can organize the results found into three major blocks.

In the first place, there are notable inequalities in entering formal employment by gender, race, and, above all, socioeconomic status. Graduates from highly educated families achieve the best occupational positions, such as non-typical occupations requiring higher education, which have the highest salaries. Male and white graduates also stand out in a predominantly female and black field, having advantages in the most lucrative occupations. These results call into question, albeit not conclusively, the "meritocratic power" thesis of higher education, as there seem to be significant differences in the destinations that the groups representing the "elite" of these courses and other graduates occupy in formal employment.

Second, the results suggest the importance of academic training and trajectory for the professional insertion of teacher education graduates. More experienced graduates with greater mastery of specific teaching skills have a higher chance of formal employment and of working in typical occupations.

A third set of findings shows that qualitative differences between fields of study matter for graduates to have different entry points into formal employment. Graduates from Mathematics and Language fields have higher chances of being formally employed, with Mathematics being the only field where privileged graduates have equal chances as their disadvantaged peers. This field also has the highest chances of graduates working in typical teaching occupations, which is the opposite situation of the Natural Sciences and Language fields.

Despite many differences regarding selection criteria for entry and career progression, the legal and institutional norms that regulate the teaching profession are closely linked to the attractiveness and retention capacity of teachers in the education system. These norms are associated with both the content required in teacher training and the competencies required, and in some cases, specific certification requirements (^{VAILLANT; MANSO 2022}). The lack of clear parameters for the teaching profession is reflected in both the definition of competencies for the position and the relative position of the professional group. As ^{Vaillant and Manso (2022)} pointed out, attracting and retaining qualified personnel in a profession that has lost prestige and status has become a crucial problem on the public agenda.

The focus on the attractiveness of the teaching profession and the differentiation between areas of knowledge within the profession itself indicate the existence of problems in the social foundation of this professional group. There are few doubts about the demands that can be made of doctors. However, even though the tasks assigned to teachers are known, intense debates persist about how to train these professionals and about the contours, limits, and framework of their role in the education system. Teachers are part of a professional group that has difficulty delimiting and controlling the specific knowledge of their work. What is the proper teaching work? There are significant variations according to the level of the education system and also between areas of study. Teaching math is different from teaching Portuguese or biology. What is the teaching/method in the chemistry class and what is chemistry? The core of each profession is established through jurisdictions (^{ABBOTT, 1988}) or areas of knowledge that a specific group controls by defining technical and social rules of belonging to the said group. Unlike doctors, teachers do not define which knowledge "belongs" to the professional group: the debate about teaching as a mission or teaching as a technique shows the degree of dissent about such an essential item in the institutional construction of a group.

This weakness can be translated into social strength, allowing for greater flexibility in professional practice and greater adaptability in the process of internal differentiation within the profession (^{FREIDSON, 1986}). Different professions tend to distinguish among three types of positions among practitioners: the most common and basic being the technical professional, responsible for the daily execution of the tasks of the occupation. A second prominent position would be occupied by academics, responsible for research, teaching and the development of the specific science/technology associated with the group. Finally, in the third position would be the professional contingent dedicated to the managerial tasks necessary for the practice and empowerment of the group and its organizations. It is not difficult to perceive the contours of this internal division of tasks within a professional group, even in the case of teachers.

It is precisely in the position of the technicians that friction/differences of attractiveness can arise between more structured professions (such as mathematicians, physicists or chemists, for example) and others that are less established, such as teachers. These frictions can translate into a choice for a non-typical teaching occupation, turning teaching into a more accessible channel for a profession that is better placed socially.

Our article has several limitations. Firstly, the available data does not contain individual information (e.g. motivations, aspirations, teaching experience), institutional information (e.g. selectivity), and contextual information (e.g. the relationship between salaries by field of study and other salaries) that are relevant to the proposed analysis. Additionally, some of the variables in the data, such as work experience at the end of graduation, which does not specify occupation, condition, duration, etc. of the work, and performance in the specific component of the Enade, lack clear connections with elements that typically matter in job selection, such as experience in the field of study or practical teaching skills. Thirdly, the choice to analyze results at the entry of graduates into the workforce allows only speculation about variations in the relationships found over time as well as on the professional trajectories of graduates. Fourthly, although we have selected an extremely relevant set of results for analyzing the professional placement of graduates, there are numerous other equally important outcomes (e.g. occupational status, job satisfaction, etc.) for future research on the subject. Finally, the methodological strategy used does not allow for the attribution of causal relationships to the correlations found.

This set of observations highlights the complexity of the relationships between different social groups and the organization of the technical division of labour (^{GRUSKY; WEEDEN, 2001}). We have identified relevant social differences in the possibilities of access and success in distinct knowledge areas. Along with the need for conceptual deepening, these data also demonstrate the existence of grey areas between these knowledge areas, which are spaces of disputes between professions, whose measurement and analysis are questionable. In this study, the use of grades obtained in the specific knowledge test of Enade limits the analysis, both due to the incongruence between the areas and the undefined and low legitimacy of competencies in each professional group (^{PISSAIA et al., 2018}).