UNIFEM Asia-Pacific and Arab States Regional Programme for Engendering Economic Governance demystifying economics and empowering women

From Margins to Mainstream
From Gender Statistics to Engendering Statistical Systems

This page is an extract from the paper titled
From Margins to Mainstream From Gender Statistics to Engendering Statistical Systems.
Download complete paper (Acrobat file, 140kb)

What are Gender Statistics?

The term Gender Statistics is now widely used to describe certain areas or compilations of statistics, particularly those presented in national collections on Women and Men in [Country].

In Asia-Pacific, UNIFEM country and regional projects during the 1990s supported the production and publication of booklets on Women and Men in eight countries: China, India, Indonesia, Pakistan, Philippines, Sri Lanka, Thailand and Vietnam. Most of these have since been brought out in new editions utilizing later data sets by national statistical offices, and similar publications have been produced at sub-national level in countries such as China and Indonesia. The Statistics Division of UN-ESCAP implemented the UNIFEM regional project covering India, Pakistan, Philippines, Sri Lanka, Thailand and Vietnam and also produced a regional publication on Women and Men in Asia-Pacific. Other donors, particularly the Swedish International Development Agency (sida) through Statistics Sweden[1], have also supported work on gender statistics.

Most sources agree that there are three primary requirements for gender statistics:

1. all statistics on individuals should be collected, collated and presented disaggregated by sex;

2. all variables and characteristics should be analyzed by and presented with sex as a primary and overall classification. This means that it is not sufficient to present just one table showing the population disaggregated by sex. All tables of individual-level data – population by education, labour force by occupation, economic activity etc – must be disaggregated by sex in addition to all other variables of interest.

3. specific efforts should be made to identify gender issues and provide data that address these.

The Statistics Sweden publication Engendering Statistics[2] also identifies two further components of gender statistics work [page 11] :

1. formulation of concepts and definitions used in data collection that adequately reflect the diversity of women and men . . . . and capture all aspects of their lives; and

2. development of data collection methods that take into account stereotypes and social and cultural factors that might produce gender biases.

However, few agencies working on gender statistics have emphasized these issues and few countries, particularly developing countries, have systematically addressed them. Most of the data that appear in publications on gender statistics were obtained by conventional data collection methods that, in most cases, remain gender blind[3].

Why do we need new statistics on gender issues?

In the past, it is men who have defined the issues that society considers sufficiently important to require data to be collected. Not surprisingly, these issues tend to be of greater importance to men and to relate to men’s traditional gender roles, particularly men’s economic roles. Even demographic data that appear to relate to women’s fertility have been collected primarily to measure the size and composition of the population in order to provide input to the measurement of the economy, national product and national income. Statistical systems are now beginning to realize that data are also needed for issues that are important to women.

Special efforts are therefore needed to identify specific gender issues that affect women and men differently.

• Some issues that may be of greater relevance to women than men. For example, gender-based violence, including domestic violence, primarily affects women and girls and is not an issue for the majority of men. Unpaid family, household and community work are primarily carried out by women. Most statistical systems have very little data about either violence against women or unpaid family, household and community work.

• Policies and programmes may affect women and men differently. Thus statistics are needed to show these differences. For example, women’s gender roles as the primary caregivers for children and families involves them in a large amount of unpaid family and household work that affects their ability to participate in other activities, such as education, paid employment and decision making. Policies and programmes may affect women differently because of this unpaid work. For example, they may not have sufficient time to benefit even from programmes that specifically target women. Statistical systems in many countries are now realizing the need for time use data to show whether and how policies and programmes might affect women and men differently.

In general, countries have been satisfied to identify one or two specific gender issues, such as violence against women or women’s unpaid labour, rather than to explore the implications of the developing gender-responsive methods of data collection, presentation or analysis in any depth.

The data currently described as Gender Statistics were collected by conventional methods that did not involve a gender perspective. The term Gender Statistics is usually restricted to the kind of social and demographic data collected in household or individual censuses and surveys. In some countries, publications on Gender Statistics also include a limited set of administrative data, largely from the education, health and justice sectors and the civil service. However, the major part of the economic data used in compiling that National Accounts that measure national income, national product and economic growth do not appear in collections of gender statistics.

Statistics are Disaggregated by Sex NOT gender

The terms “gender disaggregation” or “disaggregated by sex”, although widely used, are conceptually incorrect. This usage implies that gender and sex mean the same thing. This is very confusing because sex and gender are quite different concepts and, in particular, gender is not appropriate as a statistical variable.

Sex

Sex refers to biological differences between women and men. Sex characteristics are universal across all societies and unchangeable among individuals and societies. [Even “sex-change” operations do not change the underlying genetic determinants of sex.] Statisticians use sex as a variable when they collect data on individuals according to whether they are female or male.

Gender

Gender refers to socio-cultural differences and social relationships between women and men. In the English language, gender differences can be described as feminine for those that we associate with women or masculine for those that we associate with men[4]. However, a male can have feminine characteristics, and a female can have masculine characteristics.

Gender characteristics and stereotypes – whether an individual is described as feminine or masculine - can change:

• within a very short time – for example, a woman’s appearance might be quite feminine if she is dressed in traditional women’s clothing ready to attend a formal function, but it might be quite masculine if she is dressed in men’s overalls ready to paint a room in her house;

• differs according to the perspective of the observer. For example, one person may regard an assertive manner as “(inappropriately) masculine” in a woman, while another may regard it positively as an indication of empowerment and not describe it as masculine or feminine;

• differs by culture and between countries. For example, skirts may be generally regarded as feminine forms of dress, but in Fiji a skirt-like garment is worn by men and is part of the official uniform of the police. In some cultures, trousers are regarded as a masculine form of dress, but in South Asia saluar are part of a traditionally feminine form of dress. Nursing is generally regarded as a “feminine” occupation, but in some South Asian countries, many nurses are men.

Consequently, gender is not a very useful category for statistics: gender statistics are disaggregated by sex and not by gender.

Feminine gender characteristic or stereotypes are those associated with the female sex in terms of our expectations of what is “usual”. In practice, they are not necessarily limited to women and girls. Thus, caring for small babies is commonly regarded as a feminine “gender” role and “nurturing” as a feminine gender characteristic. However, some men also care for babies, even as primary caregivers.

Masculine gender characteristics or stereotypes are those associated with the male sex in terms of what we regard as “normal” or “usual”. However, again, in reality they are not restricted to men and boys. For example, engineering or being a pilot would generally be regarded as “masculine” occupations, but there are women engineers and pilots in many countries.

Sex characteristics

By contrast, ONLY women menstruate, or can give birth and breastfeed. These are thus female sex characteristics. ONLY men can provide semen, which is thus a male sex characteristic. Most other differences between women and men are gender differences.

Why is there confusion between sex and gender?

The confusion arises because there is a close link between sex and gender in everyday life. One of the main reasons for our interest in sex-disaggregated data is its capacity to reveal sex differences in the characteristics of women and men or girls and boys that are the result of gender differences in their lives.

For example, the sex-based division of labour in both the work force and the household arises primarily because of women’s and men’s socially determined gender roles. Although most housework and child care is done by women, males are physically capable of doing such work and some men do housework and care for their children. Similarly, although most engineers and military officers are men, women can do such work and in a number of countries women are increasingly being found in such occupations.

Engendering National Statistical Systems

The concept of engendering national statistical systems goes beyond the conventional approach to gender statistics in two ways:

1. It recognizes the need to engender data collection instruments and processes. Merely disaggregating data from conventional censuses and surveys by sex is insufficient because the data collection framework and instruments themselves are gender biased. What is required is a thorough review of the data collection framework and process from a gender perspective and an engendering of each specific data collection instrument and process.

2. It recognizes the need to identify gender issues in economic statistics. A more systematic and comprehensive approach to identifying gender issues and providing data that genuinely addresses women’s needs and priorities must focus on economic statistics. In particular, the process of data collection for the System of National Accounts (SNA) must be engendered, and the conceptual framework within which the SNA is used and economic data analyzed and utilized in policy processes must be made gender-responsive.

1. Engendering census and survey data collection processes

UNIFEM South Asia in collaboration with other UN Agencies supported the Central Statistical Organization India and the Central Bureau of Statistics Nepal in a pioneering effort to engender the 2001 round of the national censuses in those countries (click here for a write up of the Nepal experience). Based on that experience and work from some developed countries, engendering data collection processes and instruments involves, among others:

• review of the conceptual basis of data collection and the questionnaire.
  
  As part of the process of engendering the 2001 Census, Nepal established a Questionnaire and Manual Preparation Committee to review gender concerns in the data collection process.
  
  In the UK Census, an important development involved a change in thinking about women’s domestic work and child care. As a result of advocacy by women’s groups, the term “economically inactive” to describe women who were not classified as in the workforce was removed from the census form because it implied that such activities were not economic and not productive.
  
  At a more basic level, the Nepal Labour Force Survey eliminated the use of the traditional labour force questions requiring the respondent to determine whether individuals were in the workforce, thus avoiding the terminology issue completely [5]. (Indonesia had discarded the traditional approach during the 1980s). Such questions tend to lead to under-enumeration of female labour force participation because male heads of households – and sometimes women themselves – do not regard many (particularly unpaid) women’s activities as work, even when those activities would technically be considered work under the definitions being employed in the survey.

• review of coding and classification systems and terminologies.
  
  For the 2001 Census, Nepal also established an Occupation and Industry Classification Committee to review gender bias in the classifications to be used. As a result, the Committee produced a new NSCO and NSCI codes for occupation and industry that provided more detailed four digit gender-sensitive breakdowns for women and men.
  The United Kingdom developed a more gender-sensitive classification system for socio-economic status and also provided a more detailed breakdown of economic activities normally carried out by women (which had previously tended to be provided only at a rather aggregated level)[6].
  
  Like several other countries, the UK also eliminated the use of the concept “head of household”, replacing it with the term “household reference person”.

• gender training for all personnel involved in data collection, including enumerators and supervisors and the development of appropriate training modules and materials for this purpose.
  
  For example, the process of engendering the Nepal Census 2001 involved Gender Orientation Workshops for management of the Central Bureau of Statistics (CBS), followed by another for middle-level management. A public workshop for a variety of stakeholders was also held on Mainstreaming gender in 2001 Census. Gender training was also conducted for all field personnel. In addition, a workshop was held on Development of a Media Strategy and Campaigning organized with CBS, UNIFEM and UNICEF and publicity materials (a telefilm and poster) focusing on gender-specific terminologies and questions included in the census schedules.

• a deliberate policy of recruiting more women as enumerators.
  
  In some developing countries, the use of women enumerators has been found to lead to significant increases in the enumeration of women’s labour force participation. Nepal recruited women to fill 20 per cent of enumerator posts and 10 per cent of supervisor posts for the 2001 Census, a significant increase on past levels.

• review and revision of tabulations and data presentation and dissemination.
  
  In addition to reviewing the standard output of the Census, UNIFEM South Asia assisted CBS Nepal to develop an additional set of tables that specifically addressed gender issues. CBS plans to bring out a separate publication presenting these tabulations – linking, for example, sex of household head with female ownership of assets and livestock, and sex of outmigrant by various household characteristics.

The term gendered data can be used to distinguish the data generated by such engendered processes from existing “gender statistics” collected by conventional gender-blind methods.

2. Engendering data collection for the SNA

The Swedish approach to engendering statistics[7], relies on dialogues between users and producers to identify gender issues and ensure the provision of data that addresses them. However, user-producer dialogues primarily involve the national statistics office (NSO) and users of its data. This is heavily demographic and social in content because NSOs are primarily responsible for censuses and surveys. In particular, the entire framework of the National Accounts, which provides the data base for most economic policy decision making largely lies outside the scope of such user-producer dialogues and thus outside current work on gender statistics[8].

Much of the data used to prepare the National Accounts are provided by a variety of other sources, particularly administrative statistics and special surveys, which in many countries are carried out by departments such as agriculture, industry and commerce. Other producers of economic data include businesses and firms, banks and the financial system in general. Engendering the collection and use of this data opens up an entirely new area of work that has hardly been touched to date. UNIFEM, the Statistics Division of UN-ESCAP and the Central Bureau of Statistics, Indonesia, are currently supporting the Statistics Office of Timor Leste to develop the national statistical system of this new country from a gender perspective.

Among the sources of data involved in the preparation of the SNA are:

• business accounts, income taxation returns, VAT and other taxation records, other administrative data (for example, local government records on building and construction);
• balance of payments statistics, imports, exports, etc
• government finance statistics, government revenues and expenditures;
• money and banking and financial statistics;
• price statistics, used for valuation of goods and services;
• population and labour force data from censuses and household surveys;
• agricultural and industrial censuses and establishment surveys.

Despite the intimidating thickness of the official manual on the System of National Accounts 1993, the national accounts of small developing countries are often relatively simple, partly due to the lack of data. However, they are also simple because they largely cover only the monetized market economy, which forms the smaller part of the real economy of in the poorer developing countries.

This in itself is a fundamental gender issue, since it is men and the goods and services produced by men and often largely meeting men's needs that form the major part of the market economy in many poor countries. The extended production boundary of the 1993 revision of the SNA partly addresses this gender issue. However, the majority of countries have not yet implemented the extended production boundary. This is particularly true of poor countries, even although the extended boundary is of most relevance to these countries.

3. Why is it important to engender the SNA?

In describing the uses of the SNA, the official manual provides strong justification for engendering the SNA:

"The SNA is a multi-purpose system . . . designed for economic analysis, decision-taking and policy-making, whatever the industrial structure or stage of economic development of a country” [9]

Among the specific uses identified by the manual are: analysing and evaluating the performance of an economy, monitoring the behaviour of an economy in terms of major economic flows (production, household consumption, government consumption, capital formation, exports, imports, wages, profits, lending and borrowing etc), macroeconomic analysis, economic policy making and decision-taking and international comparisons. Women clearly have important interests in all of these areas - interests and priorities that may be different from those of men.

Currently, not only are women's concerns and priorities largely absent from these uses of the SNA, women's contributions to the economic production and wealth of economies are largely excluded from the data that provide the basis for preparation of national accounts. Policy decisions that exclude data on much of the roles and contributions of more than half the population cannot provide optimum results.

Engendering the SNA, as part of the process of engendering economic decision-making in general and macroeconomic decision-making in particular, must be a priority, both for women and for countries.

4. What needs to be done to engender the SNA?

This is largely new territory and we do not yet have all, or even many, of the answers to this challenging question.

• Work on the measurement and valuation of unpaid work and other uses of time use data contribute to the process of engendering both the preparation of national accounts and the conceptual frameworks within which they are used.
• New dialogues are needed between gender advocates, users of national accounts – particularly economic policy analysts and decision makers - and providers of data from sectors responsible for the administrative data used in preparation of national accounts.

In some countries, such as Indonesia, past work on gender statistics in the region rather accidentally brought into the user-producer dialogues sector staff who were meant to represent users but who were actually producers of sectoral statistics. This led to broader dialogues on the producer side, although most of the sectors represented were from social sectors – health, education, religious affairs, justice and labour.

More work is needed to bring into the dialogues those responsible for economic data such as trade statistics, finance statistics, price statistics, business and production statistics in order to begin the process of identifying gender issues and gender gaps in these areas, as well as gender biases that may arise from the ways in which such data are used.

• Collaboration with the Statistics Division of UN-ESCAP in implementing its work plan to operationalize the 1993 revision to the SNA in the region.

The UNIFEM Regional Programme for Engendering Economic Governance in Asia-Pacific and the Arab States, together with the UNIFEM Regional Offices in the region is currently working with the Statistics Division of UN-ESCAP and National Statistics Offices in a number of countries to identify ways of engendering the collection and analysis of economic, as well as social, demographic and administrative data.

Footnotes

[1] SIDA also provided technical support to the UNIFEM regional and China projects, as well as follow-up activities in China and Vietnam. The Asia Development Bank has supported follow-up activities at the provincial level in Indonesia.

[2] Birgitta Hedman, Fransesca Perucci, Pehr Sundstr`m, Engendering Statistics. A Tool for Change, Statistics Sweden, 1996.

[3] A recent exception is the national statistics offices in India and Nepal, which both implemented systematic and comprehensive approaches to engender the 2001 round of the national census.

[4] Unfortunately, many other languages are not able to make such a clear distinction between the two concepts.

[5] With increasing use of computers, there is really no need for such questions. Classification of the population by economic activity status can be easily carried out by computer during data presentation.

[6] Linda Murgatroyd, “Developing gender statistics in the UK”, Radical Statistics, No. 074, 2000

[7] Birgitta Hedman, Francesca Perucci and Perh Sundstrom, "Engendering Statistics: A Took for Change", Statistics Sweden, 1996

[8] Many NSOs are responsible for coordination of the national data base for the SNA and presenting the National Accounts and other economic data. However, the dialogue on gender and statistics within NSOs has not yet reached this area of their work.

[9] System of National Accounts 1993, Commission of the European Communities – Eurostat, IMF, OECD, UN and World Bank, 1993: 1.29.

dated: 1May2003