12.5 Statistical farm register and other frame sources for agricultural censuses and surveys#
In the past, an economic statistic programme typically included a periodic economic census that was based on area enumeration of all establishments and that provided benchmark data and frames for production and other economic surveys conducted between censuses. As noted in Chapter 12.3 — Statistical Business Register, this situation has evolved so that fewer countries now conduct an economic census. Instead, data are obtained from administrative sources and from an expanded suite of economic surveys. The frames for these surveys are derived from an SBR, which itself is based on and updated from administrative data.
It could have been expected that agricultural censuses and surveys would evolve in much the same way, in other words that fewer countries would conduct an agricultural census, that data would be obtained from administrative sources and from an expanded suite of agricultural surveys, Eventually, the frames for these surveys would be derived from a statistical farm register (SFR) based on and updated from administrative data sources. However, whilst there may have been some movements in this direction, it has been inhibited by (at least) two significant constraints.
First, there is a continued ongoing demand for data that can only be obtained from an agricultural census, or at least, a very large sample;
Second, an SFR containing a comprehensive list of farms (or agricultural holdings, as the basic unit for agricultural census and production surveys is more precisely termed) cannot be constructed from administrative sources in the way that an SBR can. The administrative sources that exist are not sufficient. Therefore, inevitably, some agricultural frames have to be constructed, at least part, by area enumeration.
Although there are no international guidelines for SFRs, there are international guidelines for conducting an agricultural census that include how to construct the census frame, as further discussed in the Chapter 12.5.2 — Frames for agricultural censuses. There is also a global strategy to improve agricultural and rural statistics, which also has information about constructing the frames for agricultural surveys, as further discussed in Chapter 12.5.3 — Methodology and guidelines for the generation of agricultural survey frames. Chapter 12.5.4 — Examples of National Practices provides examples of national practices.
12.5.1 Sources of frames#
The frames for agricultural censuses and surveys can be obtained by the following methods:
full enumeration of all agricultural holdings in the country specifically for, and part of, an agricultural census;
enumeration of agricultural holdings in a representative sample of areas specifically for an agricultural survey;
full enumeration of agricultural holdings and own account agricultural activities during the course of a population and housing census;
enumeration of agricultural holdings and own account agricultural activities in a representative sample of areas during the course of a household survey;
or from the following sources:
cadastral or other land records that identify land parcels;
an SFR that is created from data collected during an agricultural census and subsequently maintained by the NSO using any combination of the above enumeration methods and/or administrative sources and/or feedback from agricultural surveys;
an agricultural master sampling frame (MSF), comprising a representative list of units, created from an agricultural census and/or a population and housing census, maintained by the NSO, and used as a general-purpose sampling frame for selection of samples for agricultural surveys;
an SBR that is maintained by the NSO and that can provide a frame of businesses (in the formal sector) that are agricultural businesses.
In a situation where no single frame provides adequate coverage a dual frame approach may be used. Samples are selected from two frames, typically one being list based and the other area (enumeration) based. Taking care to avoid double counting of units that are in both frames, the results are combined to provide better overall coverage than could be obtained from either frame alone. The approach can be extended to three (or more) frames.
In summary, although some countries do maintain a Statistical Farm Register (SFR), SFRs do not play the same major role in agricultural survey programmes as do Statistical Business Registers (SBRs) in economic statistics programmes. Indeed, as further elaborated below, many developed NSOs, such as the Australian Bureau of Statistics and Statistics Canada, no longer maintain a separate SFR. Instead, they may rely on a frame of agricultural businesses provided by the SBR coupled with an Agricultural Master Sampling Frame (MSF) of agricultural holdings obtained by enumeration in a representative sample of areas. This explains why there are no current internationally acknowledged guidelines, principles and recommendations for SFRs and why this section is entitled Statistical farm register and other frame sources for agricultural censuses and surveys rather than simply Statistical farm register.
12.5.2 Frames for agricultural censuses#
Scope of and approach to agricultural censuses
As noted in the World Programme for the Census of Agriculture (WCA) 2020, FAO (🔗), the scope of an agricultural census varies depending on the resources available within a country, the importance of agriculture, the traditions in census organization, and national priorities. Thus, the census should be conceived as part of an integrated system of agricultural statistics. Typically, the main aim of an agricultural census is to provide basic data on the structure of the agricultural sector, such as size of holdings, land use, land tenure, etc. These are data that do not change quickly over time.
Many countries do not have sufficient resources to conduct a full agricultural census, or a suite of agricultural surveys to replace it. Thus, they collect the data from a very large sample, still referred to as an agricultural census, or they collect only a subset of data items from all holdings and collect the remaining data from a sample.
The WCA 2020 recommends a modular approach to agricultural census and survey planning, with a core module based on a complete enumeration collecting a limited set of key items, coupled with sample-based complementary, thematic surveys. This approach is in line with the Global Strategy to Improve Agricultural and Rural Statistics (🔗) (further described in Chapter 12.3.3 — Administrative data inputs for SBR construction and coverage) and the UN Guidelines on Population and Housing Censuses (🔗). To facilitate implementation, the WCA 2020 provides guidance on agricultural censuses carried out by countries in the period between 2016 and 2025. It highlights the technical and operational benefits of coordinating the population census and the agricultural census. The WCA Volume 1: Programme, concepts and definitions guidelines (🔗), as further discussed below, are in two volumes in each of which there is practical information how to create frames.
Statistical units of agricultural censuses
The statistical unit for an agricultural census is an agricultural holding and is defined as an economic unit of agricultural production under single management comprising all livestock kept and all land used wholly or partly for agricultural production purposes, without regard to title, legal form or size. Single management may be exercised by an individual or household, jointly by two or more individuals or households, by a clan or tribe, or by a juridical person such as a corporation, cooperative or government agency. The holding’s land may consist of one or more parcels, located in one or more separate areas or in one or more territorial or administrative divisions, providing the parcels share the same production means, such as labour, farm buildings, machinery or draught animals.
There are two types of agricultural holdings:
holdings in the household sector, that is, those operated by household members; and
holdings in the non-household sector, such as corporations and government institutions.
For the non-household sector an agricultural holding often, but not always, coincides with an establishment (as defined for an SBR and business statistics), depending upon the exact definitions used for each.
In most developing countries, the majority of agricultural production is in the household sector. The concept of agricultural holding is therefore closely related to the concept of household. Standards for defining a household are laid down by the UN in its guidelines for population and housing censuses as follows:
“The concept of household is based on the arrangements made by persons, individually or in groups, for providing themselves with food or other essentials for living. A household may be either (a) a one-person household, that is to say, a person who makes provision for his or her own food or other essentials for living without combining with any other person to form part of a multi-person household, or (b) a multi-person household, that is to say, a group of two or more persons living together who make common provision for food or other essentials for living. The persons in the group may pool their resources and may have a common budget; they may be related or unrelated persons or constitute a combination of persons both related and unrelated”.
For the household sector, there is usually a one-to-one correspondence between an agricultural holding and a household with own-account agricultural production activities (either for sale or for own use); in other words, all of the own-account agricultural production activities by members of a given household are usually undertaken under single management. There are two special cases where the concepts of agricultural holding and household with own account agricultural production may diverge.
If there are two or more units making up a household, such as where a married couple lives in the same dwelling as their parents, the two units may operate land independently but, as members of the same household, they make common arrangements for food, and they pool incomes.
In addition to an individual household’s agricultural production activities, a household may operate land or keep livestock jointly with another household or group of households. In this case, there are two agricultural holding units associated with the household and two sets of activities: (i) the agricultural production activities of the individual household itself; and (ii) the joint agricultural operations with the other household(s).
Thus, some countries may prefer to define an agricultural holding in the household sector to coincide with a household with own-account agricultural production. The benefits of doing this are:
identification of the holding in the agricultural census is simplified; it is no longer necessary to find out about multiple holdings within the same household;
it is in line with the practice already used in previous agricultural censuses in many countries;
the use of a common statistical unit – the household – enables the agricultural census to be more easily linked to the population census;
it facilitates the analysis of household characteristics;
if the scope of the agricultural census is expanded to include other households not engaged in own-account agricultural production, there is a common unit between agricultural production units and other households.
In defining the statistical unit for the non-household sector (corporations and government institutions), Volume 1 states that the concept of establishment (as an economic unit engaged in one main productive activity, operating in a single location) should be used.
Agricultural Census Frame Construction
Main types of agricultural census frames
There are two main types of frames for collecting agricultural statistics, a list frame and an area frame. A list frame comprises a list of agricultural holdings whereas the area frame comprises segments of land. In some cases, a multiple frame approach is used, in which part of the population (e.g., incorporated and government holdings) is covered by a list frame and the remainder (household holdings) by an area frame. For the census of agriculture, the list frame is the most common frame, and this is described below. An ideal frame is a list of all agricultural holdings, based on the operational definition of the agricultural holding adopted by the country, identifying each unit without omission or duplication and without including any units other than agricultural holdings. Such a list could be obtained through a population and housing census for the household sector, an SFR, a listing exercise or other sources. The frame for the non-household sector can also come from an SFR, an SBR, administrative records or other data sources.
Whatever method is used, special care must be taken to ensure that all agricultural holdings are included in the frame. If holdings are missing, they cannot be enumerated in the agricultural census, and the validity of the census results is compromised. This is especially significant in an integrated agricultural statistical system, as any weaknesses in the agricultural census frame are also reflected in all the surveys that follow.
Agricultural census frame from SFR
Where an SFR exists, it is a good frame for an agricultural census, provided:
it includes all agricultural holdings according to the definition established; and
it is regularly updated, to remove units that cease to operate as holdings and to add new holdings.
Along with the usual identification criteria, an SFR usually contains some basic information about each unit, such as land area, types of livestock kept, types of crops grown. This information is updated periodically. It is used for stratification when sampling techniques are used.
An SFR can be created in various ways. It may be initially created at the time of an agricultural census and regularly updated thereafter, using information from various sources. In other cases, it may be established by law as part of an administrative process and updated annually. The problem with this type of register is that the information is required for administrative purposes and may not coincide with statistical needs. The unit in the register may not correspond with the definition of the agricultural holding. For example, the register may be based on cadastral or other land records in which each parcel of land is identified rather than the holding unit. A register-based on land ownership is not entirely suitable for an agricultural census because several people in a household may manage land separately. Thus, ownership and management are not always synonymous. Furthermore, the landowner is not the land operator if the land is rented out. Frames based on business registration or licensing procedures are not always suitable as they represent what the unit is licensed to do, not necessarily what it actually does.
For the non-household sector, the SFR may be based on records from government regulatory agencies. Most countries have business registration and licensing systems. Membership information from industry associations may also be useful.
Agricultural census frame from population and housing census
Another type of frame covering the household sector of an agricultural census is one created from the population and housing census as a one-time exercise, but which is not kept up to date or maintained as an SFR afterwards. This is possible if the population and housing census include questions on agriculture that identify households involved in own-account agriculture. Although these are not the same as agricultural holdings, they provide a starting point. Based on the list an NSO can:
contact each household in the agricultural census;
ask each household about the household’s own-account agricultural production activities and the management of agricultural activities in the household, to identify each agricultural holding; and
enumerate all agricultural holdings.
The definition of the agricultural holding must be kept in mind. Where this is based on minimum size criteria such as land area, numbers of livestock, numbers of fruit trees, etc., these minimum size criteria need to be identified at some stage.
For such a frame to be useful, the agricultural census needs to be undertaken as soon as possible after the population and housing census to ensure that the list of households involved in own-account agriculture is current.
Construction of an agricultural census frame from scratch
If it is necessary to build a frame, the country is divided into suitable geographical units, referred to as enumeration areas (EAs), covering the whole in-scope national territory. Then each EA is visited, and interviews are conducted with local authorities or visits are made to households to identify all agricultural holdings in the EA. Population and housing censuses are usually conducted using EAs as the basic building block, and it is often possible for the agricultural census to piggyback on the population and housing census field system by using the same EAs and making use of maps and other field materials.
Multiple frame approach
Typically, a combination of frames is used for the agricultural census. Often, the household sector is enumerated based on the EA frame of the population and housing census, whereas a frame of agricultural holdings in the non-household sector is obtained from administrative sources. WCA 2020 Volume 2: Operational Guidelines (🔗), which elaborates the operational aspects of conducting an agricultural census, Chapter 14 - Data, Information and Knowledge Management provides detailed information on frame construction.
12.5.3 Methodology and guidelines for the generation of agricultural survey frames#
Global strategy to improve agricultural and rural statistics, 2010, FAO and World Bank
The third pillar of the Global strategy for improving agricultural and rural statistics (GSARS) is integration of agriculture into national statistical systems. Integration is needed for several reasons. A common shortcoming in both industrialized and developing countries is that data are collected by sector, using different survey frames. In addition to the NSO, there may be several other NSOs involved in the collection and analysis of agricultural, fishery and forestry data, and there is insufficient coordination. Surveys may be conducted on an ad-hoc basis with no links to a common frame or the use of georeferenced units. Thus, it may be difficult to combine data from different surveys, for example, crop and livestock production, for in-depth analysis. Integration of statistical systems can help solve many of these problems.
A very important aspect of integration is the development of an agricultural master sampling frame (MSF) and its use in a coordinated data collection program, thereby ensuring good coverage and coherent statistics. An MSF provides the basis for the selection of probability-based samples of farms and households and enables the characteristics of farms to be connected with those of households, and with land cover and land use dimensions. The challenge in developing an MSF is that it must satisfy the needs of three statistical units: agricultural holding, household, and land parcel. In addition, these three units must be linked so that, for example, household income, health, and other factors may be compared to the farm’s economic situation, and all of these to their general environmental impact. Where there is a one-to-one relationship between the three types of unit, economic, social, and environmental information can be collected from a single unit. The main difficulty in the development of MSFs occurs where there is not a one-to-one link between the agricultural holding and the household.
Regulation (EU) 2018/1091, 2018 on integrated farm statistics
The EU regulation (EU) on integrated farm statistics (🔗), which builds on the GSARS, provides direction and guidance to EU countries and is a source of guidance for others. It notes that, in order to avoid placing an unnecessary burden on agricultural holdings and national administrations, coverage thresholds should be established and that, for proper analysis of the structure of European agriculture, 98% of the utilised agricultural area and 98% of the livestock units on farms should be covered. For some Member States, these thresholds are too high for national purposes. However, the agricultural holdings below those thresholds are so small that a sample survey once per decade is sufficient to allow the estimation of their structure and the impact on production. The approach results in a reduction of cost and burden while still enabling the design of effective policy action to support and maintain small farming structures.
Handbook on Master Sampling Frames for Agricultural Statistics
In accordance with the second pillar of the GSARS, the Handbook on Master Sampling Frames for Agricultural Statistics (🔗) provides detailed guidelines and country examples for the development and construction of an MSF. The Handbook is intended as a reference document providing technical and operational guidance on various aspects of the development and use of an MSF in various country conditions, with an emphasis on developing countries. The Handbook recognizes the diversity of country situations and resources and consequently proposes various options.
Content and construction of an MSF
The population for an agricultural survey should be defined in terms of one of the three standard statistical units – agricultural holding, household or land parcel.
A comprehensive MSF is a listing of sampling units that, when associated with reporting units, provides complete coverage of the population of interest, as well as linking agricultural holdings to households and land dimensions.
An MSF may comprise:
a list of the names of farm operators obtained through an agricultural census,
a list of households derived from a population and housing census,
a list of commercial agricultural enterprises that are not associated with a household, or
a list of area units that are defined geographically; or
combinations of these.
If the list is of farm operators, the reporting unit is the holding associated with the name, and the items of interest are the land that it operates and the crops and livestock on that land, the households associated with that land, and the geo-referenced land. The list may also include commercial agricultural enterprises that are not associated with households and/or households with livestock but no land. Households that provide agricultural labour but do not operate a farm are excluded, as are households with small plots for which production falls short of a given threshold. This sort of list provides the linkage between an agricultural holding and the household associated with it. However, it is not a complete MSF because it excludes rural households.
If the list is of households, the reporting unit is the agricultural holding, and the items of interest include the land to which it is associated, the crops and livestock on that land, and the geo-referenced land. If the list has been derived from a population census, it will include all rural households, as well as those that are not linked to land but that have livestock contribute to the agricultural labour force or are simply rural non-farm households. It will not provide statistical efficiency at the first stage of sampling if the number of households used as an indicator of size when selecting PSUs is not well correlated to the items of interest such as crop areas or livestock inventories. It can become a more useful MSF if a listing of commercial agricultural enterprises is included. Alternatively, the agricultural enterprises may be used as a separate sampling frame in the multiple frame sampling context.
If the list is of segments or parcels of land, the reporting unit can be the holding associated with the land or the household(s) associated with the land having items of interest (crops and livestock on the land). Rural non-farm households within the land parcel are also reporting units. Commercial agricultural enterprises are also reporting units. This list is an area sampling frame, as well as an MSF.
12.5.4 Examples of National Practices#
EU Farm structure survey
The basic EU Farm structure survey (FSS) is carried out by all EU Member States using a common methodology every 3 or 4 years and once in ten years as a census. The basic unit is the agricultural holding. The coverage is all agricultural holdings which meet the minimum requirements set in the applicable legislation. The information collected covers land use, livestock numbers, rural development, management and farm labour input (including the age, gender and relationship to the holder of the agricultural holding).
The FSS provides comparable and representative statistics across countries and time. The data can be aggregated by different geographic levels (countries, regions, and for basic surveys also district level). They can also be arranged by size class, area status, the legal status of the holding, objective zone and farm type.
Farm Management Survey, Statistics Canada, 2017
The following description of the Farm Management Survey, Statistics Canada, 2017 (FMS) indicates that Statistics Canada does not maintain a separate SFR. For surveys such as the FMS, Statistics Canada draws the survey frame from the SBR. For other agricultural surveys, such as the Farm Produce Prices Survey (FPPS) (described in the following paragraphs) it draws the frame from the most recent Census of Agriculture.
The FMS is conducted every five years. It contributes to Agriculture and Agri-Food Canada’s work on measuring management practices on Canadian farms. The information generated from this survey helps measure management practices in the Canadian agriculture industry, address federal and provincial policy needs and support the development of effective agricultural programs. This mandatory survey focuses on both livestock and crop operations, specifically the production of dairy, beef, poultry, pig, field crops, forage crops and vegetable, fruit, berry and nut crops.
The conceptual universe for the 2017 FMS was all active farms on Statistics Canada’s Business Register. The following groups were excluded.
Active farms with agricultural sales in 2016 of less than $10,000.
Institutional farms (prisons, research stations, colleges).
Farms located on Indian reserves.
Farms for which more than 50% of the gross income in 2016 came from sales of greenhouse, sod and nursery products.
Farms included on the 2017 Greenhouse, Sod and Nursery Survey frame.
Farms without livestock inventory or crop area at the time of the 2016 Census of Agriculture.
Farms located in the Yukon, the Northwest Territories and Nunavut.
Statistics Canada Farm Product Prices Survey
The Statistics Canada Farm Product Prices Survey (FPPS) is a monthly survey that produces provincial data on average prices received by farmers for agricultural products produced and sold. The primary purpose is to provide the pricing component of the Farm Cash Receipts series, which values the agriculture sectors’ contribution to the Canadian economy. The prices are also used to calculate the Farm Product Price Index series. Agriculture and Agri-Food Canada and other federal departments use the data to develop, administer and evaluate agricultural policies and programs. Provincial departments and agencies use the prices for production and price analysis to support provincial stabilization programs.
The target population includes all Canadian agriculture operations as defined by the Census of Agriculture, as well as all marketing boards, agencies, commissions and federal and provincial government departments that collect data on producer prices or data from which prices can be calculated.