Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

An analysis of patterns of women's employment in software activities

Supposing that there are relevant opportunities for software activities coupled with a competitive local software industry offering significant volumes of employment, and that it has not been deskilled to the level of clerical work, the next dimension to be considered is the current patterns and the future prospects for women's employment in the industry.

Women's employment has tended to be associated with low skill and education requirements. Information technologies, thus far at least, do not seem to have led to any substantial alteration in this pattern, but the 'soft' side of information technologies (in technological and economic terms) that software represents may well offer a 'window' of opportunities for women. At the moment, this window may be 'ajar', although women still face difficulties in building up their career paths, on both technical and managerial grounds.

This section will analyse the patterns of women's employment in software in Brazil, and their likely prospects in the future. It is based mainly on case studies of three large Brazilian firms, undertaken from mid-1992 to early 1993. Employment figures for computing personnel were gathered, and interviews were conducted using a semi-structured questionnaire. Practical difficulties in carrying out this kind of work in Brazil limited the type of data which could be gathered, and meant that the analysis is basically qualitative. Quantitative data is used to illustrate the issues examined. For more details, see the appendix to this chapter.

A profile of the Brazilian firms studied

Software activities are hard to define, atomized and highly heterogeneous at a number of levels, and the software production process is not yet well understood. The complexity grows if one considers also software-related occupations, which tend to be spread throughout the economy. The skills required in these occupations are changing, and change is accelerating as the diffusion of information technologies progresses.

In Brazil no aggregate statistics or previous systematic analyses covering the field are available. Moreover, in our empirical work, some firms, particularly private firms, refused to divulge information. The empirical work was mainly undertaken in three large state-owned companies. Generalizations to the Brazilian market as a whole should therefore be undertaken with caution.

Company ALPHA is a computer services company, with revenues of US$ 207m and 5,037 employees. Company GAMMA is similar but smaller, with revenues of US$ 40m and 1,697 employees. Both are public computer services companies, operating as captive bureaux for Government institutions. They work as 'conventional' data processing bureaux, with a centralized structure based on mainframes, operating mainly in a 'batch' mode. However firm ALPHA, in particular, is under increasing external pressure to diversify its services, with a strong emphasis on decentralizing computer resources, coupled with the development and implementation of more advanced verticalized applications, and on their adequacy to rapidly evolving user requirement.¹⁶ BETA is a chemicals and petro-chemicals firm which is a very large information technology user. In 1991 it was the second largest employer in Brazil, with 70,000 employees. By national standards, its software activities and computing resources are remarkably decentralized.¹⁷ It is a science based firm, with a strong commitment to technological self-reliance, within what could be termed an 'engineering' mentality.

As can be seen in Table 10.2, both ALPHA and GAMMA employ large numbers of workers at the operational level of data processing, and women account for a significant share of these jobs. In firm ALPHA, around 54 per cent (2,703) of all employees are computing staff, 78 per cent (2,115) of whom work at the operational level. About 52 per cent (1,100) of jobs at this level are held by women, representing around 79 per cent of their total employment as computing staff. Their share is even higher at the base of the operational pyramid, in data entry IDE), data preparation (DP) and as production control technicians (PCT). Around 56 per cent (1,000) of jobs in these categories in GAMMA are held by women. In GAMMA, 69 per cent (1,166) of all employees are computing staff, of which 67 per cent (785) work in electronic data processing (EDP) operations. Women account for around 47 per cent (371) of all EDP jobs, but their share is higher at the base of the operational pyramid, with 54 per cent (350) of the jobs in DE, DP and PCT . Approximately 76 per cent of all computing jobs held by women are in operations.

ALPHA employs 148 programmers, 375 application analysts and 65 systems analysts, to carry out its development work.¹⁸ Women account for 32 per cent, 27 per cent and 15 per cent of these categories, respectively. GAMMA employs 28 EDP documenters, 55 programmers, 165 organization and methods analysts and 133 EDP analysts, of which 61 per cent, 35 per cent, 33 per cent and 23 per cent, respectively, are women.

Table 10.2 Educational requirements and numbers of operational and development (a) computing staff

Source: Elaborated from data supplied in 1992 by firms ALPHA, BETA and GAMMA.

(a) Figures for development staff are printed in bold text

(b) Level I education repressents the first 8 years of general education (ages 7- 15)

(c) Level II education is the subsequent 3 years of education (ages 16-18)

(d) Level III refers to tertiary graduates

(e) ALPHA requires programmers at the c level (its career top) to have a graduate degree. None of the firms said that they required analysts to have a degree in computer science' but they did require a degree plus complementary training in the computer science field.

For firm BETA, the figures in Table 10.2 refer only to computing staff located in the firm's headquarters. According to these figures, the composition of its computing staff is very different from ALPHA and GAMMA, with operational workers accounting for a negligible share of about 9 per cent (44) of total computing staff (494), and data entry being practically non-existent, whereas on the development side it has 92 programmers and 358 EDP analysts. The participation of women is even lower than in the two previous firms, at 12 per cent and 19 per cent, respectively, of operational and development jobs. According to data reported by Duarte Pinto ( 1993), the firm actually employs a total of 1,500 people in computer-related job categories, of which 326 were programmers and 562 were EDP analysts. This is only about 2 per cent of the corporation's total staff. There are 1,852 professionals located at the company's headquarters and holding graduate degrees in natural and applied sciences (e.g. geologists, geophysicists, chemists and engineers). It is striking that only 6 per cent of these, (111 jobs) are held by women, a much lower share than in computing occupations. At least in the short term, male dominance is likely to continue in software development work for such technical and scientific fields. The interviews conducted for this study, however, focused mainly on personnel involved with commercial data processing and mainframes.

Labour turnover in these firms tends to be very low, with internal labour markets. Permeability with external markets is also low, so that the opportunities for career progression inside the firm become a quite important issue. Although seniority within these firms may not be considered as a formal criterion for career promotion (or progression), it appeared to be quite relevant to the progression of computer-related occupations. Nonetheless, institutional rigidities could be observed, varying between firms and over time. There was some job mobility, not only within software development occupations, but also from the operational staff to development occupations. The latter movement was particularly visible in ALPHA and GAMMA. which have many operational employees.

The management of the production process corroborates the findings of Friedman (1986) and Friedman & Cornford (1989). At the operational level, scientific administration principles prevail, although the corresponding instruments (e.g. time and motion studies), are generally not implemented. At the level of development work, more 'flexible' managerial practices were observed, both in terms of organizational structures and mechanisms for staff control.

In contrasts to the 1970s' pattern of pools of programmers, separated from the analysts, most of the development work was organized in teams, specialized according to the type of product. This creates opportunities for master-apprentice, on-the-job learning and for job mobility. No standard structure was found for job categories, job titles or organizational roles in these firms, and job titles do not always reflect the actual fob contents. For instance, in BETA and GAMMA, the broad title of EDP analysts covers staff working with both systems and applications software, although these were clearly conceived of as suboccupations, and were in fact located in different departments. The fusion between analysts and programmers, as observed by Friedman and Greenbaum ( 1984), is taking place although the two categories retain distinct titles.

As a senior EDP analyst at BETA remarked, the conventional wisdom or 'rule of thumb' has been that two programmers are required for every analyst. In these firms, however, there are many less programmers than analysts (see Table 10.2). In ALPHA there is one programmer for three analysts, in BETA the proportion was I to 2. To say the least, employment prospects for programmers are less than dynamic.

The subcontracting of computing activities has recently become a major issue for Brazilian user corporations. The effects will be felt primarily at the bottom of the operational pyramid, where women tend to be concentrated. For instance, in 1992 the data entry work which ALPHA subcontracts was estimated to require 1,029 workers per month, almost twice the stable labour force (565) actually working in this activity. The data preparation work which is outsourced was estimated to require 113 workers, which is only about 20 per cent of the stable labour force.

Thus unskilled and semi-skilled computing jobs are tending to be placed on a flexible and unstable contractual basis. This has negative implications not only for the nature of employment on the operational side, but also for job mobility from operational to development work. The less skilled tasks of software development may also be subject to the same tendency, as can be seen in Indian software exports (Schware, 1992).

These firms employ significant numbers of skilled computing staff. The opportunities for career advancement which they offer are essentially on internal labour markets. But these companies follow, in broad terms, one of the basic principles of scientific administration - the separation of conception and management from execution activities. Their organizations have hierarchic levels (e.g. managerial and supervisory positions), and the occupations are also split into levels with a pyramid structure. Although these companies offer some scope for job mobility, it is likely that these conventional and rather rigid organizational structures inhibit both the social and technical changes necessary for the diffusion of information technologies. as well as limiting the opportunities for women to progress in their careers.

The participation of women in software occupations

The occupational structure of development work is quite complex and job fluid, with considerable room for change and mobility. Development work seems to be affected by a combination of evolving skills requirements with informal social networks and high degrees of autonomy at the organizational level.

The evidence from these firms is that women's employment is concentrated at the base of the pyramid of computing jobs. Women in operational work tend to be found mainly in data entry, data preparation and as production control technicians. These jobs involve clerical-like tasks, with low skills and low education requirements. To use Cockburn's terminology, these jobs require the 'know what' kind of technical skills, with very few chances of career advancement. She also remarked that women generally tend to be reluctant to engage in technical careers:

It does not spring from inadequacy or lack of interest but from a keen perception of the costs involved: isolation, discomfort, harassment and, often, wasted time and energy.... Technology and the relations of technical work have to change before most women will choose to engage with them.

(Cockburn, 1985: p. 13)

The data shown in Tables 10.3 and 10.4, for firms ALPHA and BETA, does show that women are the 'occupational losers' in computing jobs. For instance, although experience and time in the company were said to be important criteria for promotion, women tend to stay longer in these firms and to concentrate in the lower levels of each job category, an indicator that their careers advance more slowly than those of men.

Women's career progression: rising to challenges

Although the statistical picture found in the case studies does not paint a rosy picture for women, the qualitative interviews with seventeen women working in software occupations may be grounds for a more optimistic perspective. In most cases they had found room for career progression, including moves from operational to development activities. Far from being reluctant, their professional attitudes were determined and competitive.

In spite of being in relatively disadvantageous positions, the vast majority of the women interviewed valued the job stability and fringe benefits such as subsidized medical assistance and children's schooling, as well as the rather flexible working hours of the occupation.

Table 10.3 Firm ALPHA. computing staff, distribution of job categories by sex and employment time

Source: Data supplied in 1992 by firm ALPHA

Table 10.4 Firm BETA, computing staff. distribution of job categories by sex and employment time

Source: Elaborated from data supplied in 1992 by firm BETA

In contrast to Friedman's findings in the USA (Friedman, 1986), operational activities have served as a significant route for women entering computing occupations. Of the seventeen women interviewed, eight had taken this route to building up their career paths, albeit with great difficulty. As L.U., an EDP analyst at GAMMA who started in data preparation remarks, 'Either you join the company in a good position or you have to endure a lot of pain'.

The pioneers in the early 1960s had to face explicit prejudice. Three analysts at ALPHA, now in their mid to late fifties, had started as key punch operators (I.S. and M.L.) or in data preparation (A.M.). All have children and have raised their families as the sole 'breadwinners' while working, sometimes in two different places, and at the same time earned their university degrees through evening courses. In the mid 1960s they were offered a first opportunity to attend courses in computer programming and, at the same time, were explicitly discouraged from pursuing it. As I.S. puts it:

When programming emerged in Brazil as a profession, it brought a pattern established abroad. Then, it was clearly set as male's work, whereas data entry jobs were regarded as appropriate for women. At that time, men working in the latter were also discriminated against. People often wondered whether they were gay or physically disabled.

A.M. also recalls her first experience with programming, quite illustrative of the kind of resistance these women had to endure:

In my first chance to attend a programming course, the teacher asked me what the hell I was doing there. My stay was conditioned on remaining silent so as not to disturb my male colleagues. I was allowed to develop the final project. My programme was the last to be tested and the only one that worked with no problems. Then the teacher recognized that, in spite of being a woman, I was capable of programming, inviting me to undertake a more advanced course. For a woman to succeed, she had to work extra hard. I believe that when the labour market began to expand the prejudice against women diminished.

Contrary to Cockburn's hypothesis, women in Brazil have not been reluctant to follow these pioneers in pursuing computing careers. In fact most of them seem to have learned to play the game according to the males' rules, that is, to be assertive and to compete fiercely to get what they want. The professional behaviour of C.L., a senior EDP analyst at BETA, is illustrative of such an attitude. She joined the firm in 1980. Very early in her career she was chosen to be a project leader. For the last 6 years she has been a junior manager of development teams, now working with the development of systems to support managerial decision-making. C.L. is very assertive in defining her entrepreneurial attitude, 'Women who gain a bit of prominence are "phallic", that is, they compete aggressively like men. They are courageous, do not escape from conflict but, quite on the contrary, face it upfront.'

The division of work in the domestic sphere does not seem to have gone through any radical changes either. Ten of the women interviewed have been raising children. A few of them said they had delayed having children because they were afraid of being devalued professionally. Those who were married or had partners tended to classify them as 'very supportive', although the women continued to carry the main responsibility for housework and childcare. Nevertheless, the vast majority of these women seem to cope very well with their professional and domestic lives. The sad side of these rather 'successful' stories is that, in many cases, the responsibility for 'rocking the cradle' is, in fact, taken over by other women working as full-time house servants, because men generally do not share these responsibilities.

Some of these women, either intuitively or consciously, are initiating changes in the hierarchic social relations at work. For instance, six out of the eight women interviewed with managerial experience had been adopting a participatory style, involving cooperative work and democratic relations. The initiative of L.U., an EDP analyst at GAMMA, is remarkable. She was assigned to supervise a group of data entry personnel at a client location. She encouraged them to work as a team and to take the decisions of how and when to carry out their tasks. L.U. notes that:

People became highly motivated to work. Women, in particular, loved to participate in the decisions about their own work and the flexibility they got. Only one man did not adjust to the scheme. He was used to working individually and complained to the operational director at GAMMA headquarters. My experience is that women are more discriminated againt at the operational level than in software development work.

Beyond the mainframe: a future for women?

In contrast to the conventional labour process view, an important finding from my research was that, in both operational activities and the more 'noble' software development work, occupations have not been sex-typed. The only areas which seem to remain the realm of men are those in close contact with mainframes.

In the literature it is generally argued that the primary allegiance of computing staff has been to the technology itself, not the organizations that employ them. This appears to be quite applicable to personnel such as systems analysts, whose technical and social prestige have been supported mainly by the hardware manufacturers, which guarantee a stable environment for their career progression by producing a continuous flow of new generations of computers, within a stable framework of upwards compatibility (as pioneered by IBM in the 1960s).

A.L., an EDP analyst at BETA, who had worked for five years as a systems analyst in an end-user department, gives some evidence of the kind of environment that surrounds these hi-tech machines:

I was assigned to give support to the operating system, very close to the machine, within a social climate permeated by machismo. Men think of themselves so highly, that they believe they are indispensable. The machine is at the centre, while users are looked down upon as rather 'dumb'. Besides being left behind, the work was not that attractive to me to the point of continuing to endure such a macho environment. I managed to be transferred to a group providing support to users, as soon as it was created in my department. This is a promising area: the future is one of increasing integration with users, which will alter the professional profile of software development labour.

Friedman identifies areas involving more communications, interface with users, and client support as likely to be conducive for women to develop their expertise. Although these areas have become central to software evolution since the 1980s, he notes that there could be a danger of segregating women in them, under the label of 'feminine-type' occupations. Despite the danger, this is a highly valued type of work and a very promising scenario for women. Interaction with users is not only a promising area for women, but also a core component of technical progress in software. In fact, user-producer interactions are central to the evolution of technologies, particularly, as Lundvall (1988) noted, in the case of rapidly changing and complex products. The quality of exchange, with the development of direct cooperation links based on mutual trust, is central. However only in the 1980s has this activity been brought to the fore in software, with increasing concern about the effectiveness of the systems being designed to users' needs. Brooks claimed that:

The hardest single part of building a software system is deciding precisely what to build. No other part of the conceptual work is as difficult as establishing the detailed technical requirements, including all the interfaces to people, to machines, and to other software systems.... Therefore, the most important function that the software builder performs for the client is the iterative extraction and refinement of the product requirements. For the truth is, the client does not know what he wants.... Complex software systems are, moreover, things that act, that move, that work. The dynamics of that action are hard to imagine. So in planning any software-design activity, it is necessary to allow for an extensive iteration between the client and the designer as part of the system definition.

(Brooks, 1987: p. 16)

A.C. and E.C., two senior EDP analysts at BETA, share a rather more radical view: users could even assume leadership of the process. As A.C., who integrates an experimental team of business analysts, asserts:

I believe that computing experts tend to keep a distance from users. In the process of identifying and specifying the systems' requirements of new applications, users should participate actively in the development team and, eventually, lead the process.

This is also the approach advocated by the users' support groups of BETA's Service of Informatics Resources in Rio. Their aim is to endow end-users with the technical capacity to be self-sufficient to develop their own applications. Implicit in this radical approach is the socialization of technical knowledge, which has potential to diminish the social power that computing experts have been able to concentrate in their hands in the past decades.

E.C., one of the leading figures in this way of thinking, notes that:

In many national meetings about the function of technical support to users, I observed a significant participation of women ahead of this process. We tend to have an ability to establish empathy, as a core element for generating relations based upon personal credibility and emotion.

Contents - Previous - Next