FAO STATISTICAL DEVELOPMENT SERIES
PROGRAMME FOR THE WORLD CENSUS OF AGRICULTURE 2000
Food and Agriculture Organization of the United Nations
Statistics for food and agriculture are very important. Detailed analysis, based on objective, quantitative data, of the structure and performance of the food, agricultural and rural sectors is necessary in formulating sector policies, and preparing, monitoring and evaluating development plans and projects relating to food, agriculture, agrarian reform and rural development.
This chapter provides an historical background on the world census of agriculture programme, emphasizes the need for a census of agriculture in an integrated statistical programme, and outlines the basic characteristics of the Programme for the World Census of Agriculture 2000 (WCA 2000).
1.1 By April 1995, 120 countries are expected to have completed, or to be in the process of conducting, an agricultural census for the period 1986-1995 covered by the Programme for the 1990 World Census of Agriculture.
1.2 The Programme for the WCA 2000 will be presented at the 28th FAO Conference to be held in Rome in October 1995. This Programme has been developed taking into account experiences and developments obtained from many national and international experts, agencies and meetings, mainly the 4th session of the FAO/IASI Working group on Agriculture held in 1993 in Buenos-Aires, the 14th session of the African Commission on Agricultural Statistics held in 1994 in Dakar and the 15th session of the Asian and Pacific Commission on Agricultural Statistics held the same year in Manila.
1.3 The Programme for the WCA 2000 recognizes the wide range in the current status of economic and statistical development in different countries, and encourages countries to develop and implement an agricultural census tailored to their unique situation. Nevertheless, the Programme encourages standardization of a minimum core programme for the agricultural census and a complementary relationship between the agricultural census and more frequent food and agricultural sample surveys, all of which should have a recognized role within the overall national survey programme. FAO has developed a document to complement this Programme, named Taking Agricultural Censuses which provides technical guidelines.
1.4 The Programme for the WCA 2000 is the eighth decennial Programme. The first and second Programmes were sponsored by the International Institute of Agriculture (IIA) in 1930 and 1940, respectively. The subsequent five Programmes - centred on 1950, 1960, 1970, 1980 and 1990 - were promoted by FAO which assumed the responsibilities of the IIA following its dissolution in 1946.
1.5 The present Programme, like the five previous Programmes, is based on Article 1 of FAO's Constitution, the first paragraph of which reads: "The Organization shall collect, analyze, interpret and disseminate information relating to nutrition, food and agriculture."
1.6 World census of agriculture should not be interpreted as a census carried out simultaneously in all countries during a given year. Such an attempt was made in the 1930 Programme by asking countries to carry out their national censuses of agriculture during 1929 if located in the northern hemisphere, and during 1930 if located in the southern hemisphere. The objective was to obtain global data referring to the same time period. A similar request was made in the 1940 Programme. However, because of difficulties countries experienced in taking the census in the same year, and also because the population census was proposed to be taken in years ending with 0, this requestwas later relaxed. For example, the 1990 Programme covered national censuses taken during 19861995.
1.7 The 1930 Programme was promoted at a time when there was a large gap in agricultural information and data sources for agricultural statistics were not well organized, even in developed countries. The agricultural census was expected to help fill this gap. However, due to its wide scope and emphasis on agricultural output, many countries experienced difficulties in taking their first national agricultural census. Adequate resources for maintaining a large field staff were not easily obtained, their recruitment and training were major concerns when professional staff were limited in number, and completing long questionnaires fatigued both enumerators and respondents. Ensuring data quality was a major problem; even if this was achieved data processing in the pre-computer era was excessively timeconsuming leading to belated release of results. All-in-all the whole operation proved to be beyond many countries' capacities.
1.8 Taking these experiences into account, the 1950 Programme proposed a more restricted scope, concentrating on structural aspects of agriculture. The 1960 to 1990 Programmes gradually reexpanded the census scope while keeping agricultural structure as the principal theme. Some countries again experienced difficulties as the scope expanded although others conducted agricultural censuses with an even wider scope.
Census of agriculture in an integrated statistical programme
1.9 Not all food and agricultural statistics can be collected through a decennial census of agriculture. Most obviously, there is a need for annual estimates of production. As early as 1966, attention was drawn to establishing an integrated set of data sources with the census of agriculture as an important component (FAO 1968).
1.10 The 1989 World Conference on Agrarian Reform and Rural Development (WCARRD), urged countries to develop socioeconomic indicators for monitoring and evaluating programmes on agrarian reform and rural development, thus placing even greater demands on national statistical capabilities. The need for a complete system of food and agricultural statistics, with reference to various data sources, is stressed in FAO publications. Primary data sources include the agricultural census, population census, industrial census and sample surveys such as those on agricultural production, livestock, farm management, postharvest losses, food consumption, household income/expenditure, and labour force. Administrative records, local community records, technological research and international statistical publications are cited as secondary data sources. These and other sources, some beyond the domain of agricultural statistics, contribute to the data pool needed for a national information system on food and agriculture. Emphasis in these publications focuses on linkages between these data sources, with the following aims:
1.11 Continued emphasis is placed on the need for countries to prepare a multiyear plan for a sequential statistical programme of activities relating to data collection, processing and analysis, and to allocate adequate resources in a balanced manner for the components of the programme. The totality of these activities and the agencies conducting them is referred to as the integrated statistical programme component of a national information system for decisionmaking in food, agriculture and rural development. Thus, a programme for a national census of agriculture should be conducted during 1996-2005 within the above framework using this Programme for the WCA 2000 as a guideline.
1.12 Planning and implementing a sequence of coherent statistical activities require an efficient organization, trained personnel at various levels, and secured budgetary allocations over a period ofyears. Efficient organization implies strong cooperation between both users and producers of statistics. As different statistical activities are not necessarily all under the jurisdiction of a single government institution, establishing coordination among various agencies is sometimes difficult because they may have different mandates regarding the purpose, scope and timing of their work. Countries which are still experiencing a shortage of trained statistical personnel and/or insufficient funds for statistical development will require time to achieve an integrated statistical programme. Nevertheless, it is recommended that all statistical development efforts should be oriented toward the longterm goal of establishing such a programme that will eventually provide a continuous flow of timely, accurate data covering all relevant aspects of food, agricultural and rural development. Needless to add that those few countries that have a programme which delivers the information needs efficiently will find the requirements of WCA 2000 incorporated within their existing arrangements.
1.13 A census of agriculture, due to its decennial frequency, is best suited to collecting data on various characteristics relating to agricultural holdings that change only slowly over time. This limits the desirable scope of the agricultural census to the number and size distribution of agricultural holdings by type of enterprise. These attributes are referred to as the structure of the agricultural production sector. Types of enterprise included in a census of agriculture are generally related to the resource base, e.g. agroecological zones, technology such as power sources, machinery and equipment, land improvements, croplivestock rotations and general source of soil nutrient improvements such as organic or chemical fertilizers. Other structural classifications may relate to the educational level of the farmer and farm labour inputs, the legal status of the holder including tenure arrangements, and other social and demographic characteristics of holders.
1.14 Structural data will not allow analysis of the performance of the agriculture sector - this requires data on prices and quantities of inputs and outputs, enterprise costs and returns and net farm income. Data on performance and complementary data on variables such as food prices, consumption and nutrition which change rapidly over time can best be collected through frequent sample surveys.
1.15 A census of agriculture is distinguished from sample surveys by its extensive, even if not total coverage of holdings. When a complete enumeration is possible, a census of agriculture provides a frame for later sample surveys to provide the performance data. The agricultural census also provides structural data for individual small areas (communities, administrative units and agroecological zones) which are needed in the preparation of plans and policies for rural development. The census is the only basic data source for the structure of specialized holdings (such as those equipped with greenhouses), and again, provides a frame for sample surveys of such specialized features. The agricultural census is also useful in identifying disadvantaged groups (such as subsistence farmers and female holders) who need to be assessed separately in policy formulation in order to ensure their levels of living are improved.
1.16 Two points need to be emphasized when considering the scope of the agricultural census. First, the census involves collecting, processing and analyzing data from a very large number of holdings so the coverage for each holding must be kept simple and succinct. Second, since the census statistical unit is the agricultural holding, structural data related to other units, such as agricultural service establishments or seasonal and occasional workers, should be left to other data sources.
Basic characteristics of the Programme for the WCA 2000
The Programme for the WCA 2000 has three basic characteristics:
1.17 First, uniformity in concepts, definitions and classifications with those from other data sources. To the extent possible, these criteria used in the census are compatible with those recommended by other UN organizations. For example, the definitions adopted in this document for such concepts as household and occupation conform with those recommended by the UN Statistical Division.
1.18 Second, the data coverage within the agricultural census (as detailed in Chapter 4) should be limited. Given the scale, resource requirement, field management and enumeration skills and data processing implications this limitation is essential if the success of the census is to be ensured.
1.19 Third, in relation to the previous WCA Programmes, changes have been kept to a minimum. Some changes are necessary to reflect current priorities and to improve the presentation. New to the WCA 2000 are basic items that address gender and environmental issues (see Appendix 1) and reference to the need for geo-referencing data