January 2005

Gender concerns in agricultural census in Africa

by Diana Tempelman,
Senior Officer, Gender and Development

and Naman Keita,
Senior Statistician
FAO Regional Office for Africa

This paper was presented at the third International Conference on Agricultural Statistics, “Measuring Sustainable Agricultural Indicators” 2 – 4 November 2004, Cancun, Mexico

1. Introduction

There is a general lack of gender specific data relevant to the agricultural sector. In many countries, past policies for agricultural development often narrowly aimed at product growth, overlooking the importance of human resources as well as the social and welfare aspects of development. This may well have contributed to the weakness of data related to these aspects in agricultural statistics until the late eighties. The “Rome Declaration” adopted by the World Conference on Food Security, November 1996, acknowledged the fundamental contribution of women in achieving sustainable food security for all and in this context, recommended improving the collection, dissemination and use of gender-disaggregated data in agriculture, fisheries, forestry and rural development. However, there are other reasons for increased demand for socio-economic and gender disaggregated data, notably in support of:

• governments’ changing role from central decision-maker to facilitator of involvement of decentralised / sub-national bodies and the private sector in the planning and implementation of agricultural development; and
• the planning of responses to relatively recent emergencies, like increasing poverty, food insecurity, feminisation of the agricultural sector and the HIV/AIDS epidemic.

2. Improved data collection regarding human and socio-economic factors

Collecting gender sensitive data goes beyond the simple disaggregation of data by sex of Heads of Holdings/Households or the agricultural population: gender differences refer to socio-cultural and economical factors which are determined by the society and the period in which a person lives, while sex describes biological differences which do not alter over time and place. Gender-sensitive data thus attempt to reflect the diverse and differentiated situation of men and women and their specific contributions to the agrarian economy. This means that standard statistical concepts, definitions and methods used in agricultural statistics need to be checked on gender biases and where necessary rectified.

In this regard FAO supported Member Countries in analysing possible gender biases in:

• The statistical unit used in agricultural surveys and censuses: the agricultural holding. In most African countries, a one-to-one correspondence is usually assumed between a household and a holding whereby the household is used to identify the holding. This potentially contributes to the under-estimation of the number of (sub) holdings belonging to one household, usually run by women farmers managing their own sub-production units within the male headed agricultural holding. During the 2000 round of the World Census of Agriculture, several African countries have tried to rectify this phenomenon through the adoption of the concept of sub-holder, introduced under different names: head of plot or field, plot manager or “responsable de parcelle” in French speaking countries. Please refer to table 1 to see the impact of the introduction of the sub-holder concept on information concerning the distribution of areas cultivated by crop.





• The concept of Head of Household. In Africa a tendency has been observed that the eldest male household member (whether usually present or not), is automatically recorded as the 'head of household'. Socio-cultural perceptions may well have contributed to the following low figures of female-headed households in: Guinea (2 %), Tunisia (6.7 %) and Benin & Burkina Faso (10 % each). Efforts have been made to remedy these culturally prompted responses with interviewer training and applying the rule of a minimum presence of 6 months per year of any household member.



• Agricultural labour - Labour is a crucial factor in agricultural holdings, especially so in the smaller ones. The difficulty concerning women's labour is to distinguish between their productive and household activities. The 2000 WCA has seen various attempts in Africa to improve data collection regarding agricultural labour looking into family / hired labour, remuneration in cash or kind and the use of mutual assistance groups.

3. Results from selected countries

Experiences show that:

• it is crucial to include the need for gender specific data in the objective of a National agricultural census, thus justifying and ensuring all further actions;
• there is no immediate need for new statistics methods to integrate gender concerns (Senegal), but there is a clear call for fine-tuning the concept of Sub-holder, better training of census staff and a complete revision of the analytical framework of statistical data collected (Mozambique and Namibia);
• the Sub-holder concept is one of most important innovations tried during the 2000 WCA, permitting extensive cross tabulation of production and socio-demographic factors, thus allowing for in-depth analysis of intra-household gender and age based differences in agricultural production;
• there is still a wide lack of familiarity among data producers and users about what ‘gender’ concepts actually entail, which contributes to an overall weak demand from users of gender specific data.

In conclusion,
First, there is an urgent need for strengthening of the capacities of:

• Statisticians regarding the gender explicit analysis of agricultural data; and of
• Development planners to use available data in development planning in general and gender disaggregated data for gender specific planning. This would increase demand for gender specific data and would provide for the necessary feedback to the statisticians on the usefulness of the data collected.

Second, the importance of the availability of sub-national level data can not be stressed enough, as these illustrate regional differences in overall agricultural performance, as well as gender differences resulting, among others, from male dominated rural out-migration. Only such data show the extend of the feminisation of the agricultural sector, whereby female – male ratios in selected regions may have changed to as much as four women for every man (4:1), often found among the most productive age groups of 20 to 45 years old.

Please refer to figure 1 providing agricultural population data for Guinea and tables 2 and 3 providing the same for Cameroon, clearly showing the extent of the feminisation of the agricultural sector in selected regions of these countries.