Impact of COVID-19 pandemic on censuses of agriculture. Global review

. The Food and Agriculture Organization of the United Nations (FAO) coordinates the World Programme for the Census of Agriculture 2020 (WCA 2020). A census of agriculture is a nation-wide statistical operation, undertaken at least once every ten years, that collects data on the structure of agriculture at farm level. As a part of its monitoring role, the FAO receives regular updates from member countries on the effects of the COVID-19 pandemic on the ongoing agricultural census activities. This paper assesses the extent to which agricultural censuses were affected by the COVID-19 pandemic and proposes best practices for minimising its impact. The latest review of censuses in 155 countries and territories shows that COVID-19 has affected censuses of agriculture both in developed and developing countries. However, some countries coped with the pandemic better than others. Prior-to-the-pandemic improvements in national statistical systems, a wide range of information and communication technology (ICT) solutions, and the use of alternative data collection methods and administrative registers helped these countries to reduce significantly their reliance on face-to-face interactions, and instead made them favour online training and remote data collection. This has a potential to further accelerate the pace of innovation in census-taking. For other countries, the experience underscored the need to improve the ICT infrastructure, diversify data collection methods, adopt e-learning for training, and start using administrative registers in future censuses.


Introduction
The Food and Agriculture Organization of the United Nations (FAO) coordinates the World Programme for the Census of Agriculture 2020 (WCA 2020), which supports national censuses of agriculture conducted during the 2016-2025 round (FAO, 2015).Since the onset of the pandemic in 2020, the FAO has conducted several assessments of the impact of COVID-19 on censuses of agriculture (Castano, 2020a(Castano, , 2020b(Castano, , 2020c(Castano, , 2022a)).For each assessment, a short questionnaire was sent to national agricultural census authorities in the member countries.The collected information was supplemented by the consultations with the FAO experts assisting censuses in some of these countries.This paper presents the findings of the assessment undertaken in May 2022, which involved 155 countries (Castano, 2022b).
The purpose of this paper is to appraise the extent to which censuses of agriculture were affected by the COVID-19 pandemic, and to identify best practices for minimising this impact.
As most countries went into lockdown in the early 2020, disruptions begun occurring with regard to plans and preparatory activities related to agricultural censuses.The level of disruption varied depending on the stage at which the censuses were, with different repercussions experienced at the planning (i.e.staffing, procurement, preparation of frames, questionnaires), fieldwork (field training and enumeration) and data processing/analysis stages.These activities generally involve gatherings and face-to-face interaction.

Status of the national censuses of agriculture
The census of agriculture is, like the population and housing census, a key source of granular data, in this case on the structure of the agriculture sector.Structural census data refer to aspects of agriculture that change relatively slowly over time, such as the size of the holdings, land use, crop areas, livestock numbers and agricultural inputs collected at the farm level.National censuses of agriculture should be conducted at least once every ten years.Results are used for agricultural planning and building sampling frames for sample surveys.
The reference period for a census of agriculture is the agricultural year, i.e. the cropping seasons within one year.Therefore countries carefully schedule agricultural census activities to ensure that crop and livestock data are collected at the right time. 1 A delay in census activities can be critical and may result in the postponement of the enumeration by a full year if the agricultural season is missed.
Map shows the status of censuses of agriculture in 204 countries and territories.It shows 155 countries with agriculture censuses or plans to conduct one, while the remaining countries have no firm plans or no information is available.Approximately 60% of these 155 countries reported that census activities were either delayed, postponed or suspended.
1 Some countries have more than one cropping season and field operations may be better carried out at certain periods of the year for operational reasons (transportation, heavy rains, farmers' availability, etc.). Figure 1 focuses on 155 countries with censuses of agriculture that were reported as planned, underway or completed, expressed in percentages.The status of these censuses is described either as postponed (28% of countries), delayed (22%), completed with delays (8%) or suspended (2%). 2 The remaining 40% of the censuses were on schedule (21%), completed before the pandemic (7%), or are at early planning stages (12%).

Censuses of agriculture affected by the COVID-19 pandemic
As mentioned earlier, 60% of censuses of agriculture were either postponed, delayed, completed with delays or suspended.COVID-19 measures had a clear and direct impact on agricultural census activities that involved gatherings and human contact, such as the recruitment of field staff, training, listing operations, data collection, etc.In addition, the reduction of businesses (including restaurants and hotels), transport and other support services prevented the normal implementation of field operations.
A total of 43 countries (28% of the analysed countries) were forced to postpone their censuses.Figure 2 shows that census activities were postponed in Africa (37% of countries with postponed censuses), the Americas (21%), Asia (19%) and Europe (16%).
Censuses in 34 countries (22% of the countries) were delayed.Figure 3 shows that most of the delays in census activities were reported in Europe (32% of 2 Postponed: re-scheduled to a later date; delayed: delayed but still ongoing.countries with delays), followed by Africa (29%), the Americas (18%), Asia (12%) and Oceania (9%).
Thirteen countries managed to complete their censuses with delays (8% of the countries).Figure 4 shows that most of the censuses completed with delays were reported in Asia and Europe (31% of all censuses in each of these continents), followed by the Americas and Oceania (15% each).
Three countries (2% of all the analysed countries) had to suspend their censuses.These were: Mauritania and Egypt in Africa, and Papua New Guinea in Oceania.Countries which had to delay, postpone or suspend their censuses of agriculture reported several concerns regarding the resumption and completion of these massive operations, including: • higher census cost than planned (fixed costs continue despite suspensions) and unforeseen expenses, such as personal health protection equipment, revision of questionnaires and manuals, relaunching of training and publicity campaigns, etc.; • difficulties in continuing the financing of census activities once they were resumed, as governments faced other post-COVID-19 priorities; • unwanted changes in the census reference periods or problems with recalling accurate answers by producers if fieldwork was postponed, but reference periods remained unchanged (recall bias); • possible data quality issues as census data might not reflect the new reality of the post-COVID-19 agricultural sector; sampling frames might also become outdated; • delayed population censuses forced, in turn, the postponement of agricultural censuses in countries that depend on the former to establish the list of agricultural households, or those which can only organise one census at a time (Castano & Cara, 2019).

Censuses of agriculture unaffected by the COVID-19 pandemic
As illustrated by Map, censuses in 62 countries (40% of the analysed countries) were unaffected by the COVID-19 pandemic.More specifically, these were the countries that coped with the crisis well and continued activities on schedule, those which completed their censuses before the pandemic, or the ones which were still at early planning stages when the pandemic started.
Half of them (32 countries), or 21% of ongoing censuses of agriculture were reported on schedule (Figure 5).Most of these countries were able to weather the crisis and continue with major census activities.More specifically, these were 14 European countries, four French territories, the USA and five US territories, Mexico, Kiribati, Tonga, New Zealand, Nauru, Palau, Republic of Korea and Qatar.Censuses in these countries were either completed or uneventfully underway in 2022. 3Most of them had developed statistical systems, solid information and communication technology (ICT), and used administrative records 4 .This means that data collection could continue using remote methods, such as Computerassisted Web Interviewing (CAWI), Computer-assisted Telephone Interviewing (CATI), post, or by sourcing data from administrative registers.
In the case of the Pacific islands such as Kiribati, Nauru, Palau and Tonga, censuses were unaffected because these countries managed to stave off the coronavirus pandemic, reporting zero COVID-19 cases as of May 2022.
Figure 6 shows that other 19 countries or 12% of the censuses had not been affected by the pandemic by mid-May 2022.Most of these countries were in Asia (48% of censuses not yet affected), Africa (26%), and the Americas (21%).These 3 By mid-May 2022, five countries completed their censuses of agriculture during the pandemic: Poland, Estonia, the Netherlands, Republic of Korea and Kiribati. 4Administrative registers are databases held by the government, collected and used for the purposes of administering taxes, benefits (e.g.agricultural subsidies) or services (e.g.extension).Finally, 11 countries5 (7% of the analysed countries) were not affected at all by the pandemic because they had completed their agricultural censuses before the crisis struck. 6s discussed above, some countries coped with the pandemic better than others, and continued their census operations largely unaffected.Apart from the Pacific islands, where, as mentioned before, there were no reported COVID-19 cases until May 2022, most of them had developed statistical systems and reliable ICT, and also used administrative records for some of their census data needs.
In these countries, final census preparation activities continued through teleworking and e-learning, whereas data collection was implemented using several other data sources such as administrative registers and remote data collection methods, as initially planned.The latter methods involved CAWI, CATI and post (mail-out/mail-back).For example, countries such as Belgium, the Czech Republic, Estonia, France, Lithuania, and Republic of Korea curtailed face-to-face interviews and increased the use of CAWI, CATI and/or post (Eurostat, 2021).Other countries, including Iceland, Luxembourg and the Netherlands, sourced their census data mainly from administrative registers.In some cases, the pandemic negatively affected farmers' willingness to cooperate and provide all the required information.This slowed down the updating of administrative registers such as the EU Integrated Administration and Control System (IACS) 7 and national administrative registers, which are the critical sources of census data.Furthermore, the non-response rate is expected to be higher than in past censuses.

Conclusions
The COVID-19 pandemic affected most censuses of agriculture in both developed and developing countries.Thirty-two countries and territories coped better than others with the challenges posed by the crisis, helped by well-developed statistical systems, good ICT solutions and administrative registers.
Some valuable lessons can be drawn from these experiences, such as: • the key role of e-learning and online courses for trainers, supervisors and enumerators; • the importance of stable remote access to systems of the census ICT infrastructure (e.g.databases, digital questionnaires, processing systems); • the need to reduce the reliance on face-to-face interviews and to introduce alternative remote data collection methods, such as CAWI, CATI, e-mail, and traditional post (Castano, 2018).It is worth mentioning that the use of such methods was already planned by countries successful census-wise, although the pandemic shifted the balance more decisively in their favour; • the usefulness of administrative registers to meet some census data needs and avoid unnecessary duplication of efforts to collect data that are already available (FAO, 2018).The capacity of countries successful census-wise to collect information from a wide array of sources did not happen overnight.Efforts to modernise and streamline census methods and processes had already been made in the previous census round.Earlier innovations made these countries better prepared to meet the challenges posed by COVID-19.These countries were able to devote more efforts to promoting the use of CAWI, CATI, e-mail, and post interviews.The COVID-19 experience validated these efforts and will -to all probability -further accelerate the pace of innovation, by more extensive collection of remote data, the use of big data and geospatial information, and increased use of administrative registers.
7 IACS is a database administered at national level on the basis of the EU legislation.It centralises data on agricultural subsidies paid by the European Union in each member state.
For other countries, the COVID-19 crisis was a wake-up call to diversify their census data collection and training methods in future rounds.Some countries will take longer to switch to modern technologies due to poor infrastructure, high cost of access and lack of necessary skills.Some others may be able to leapfrog.The use of CATI and CAWI methods requires access to reliable and well-developed national telecommunication infrastructures.Some remote rural areas may not have access to broadband Internet.Mailed questionnaires might not be feasible in areas where postal boxes are far from respondents' homes.Furthermore, the postal service may not work adequately in times of a pandemic.
Other challenges could be related to education and literacy levels.Computer-and web-illiterate respondents might be reluctant or unable to use CAWI.Nonetheless, efforts may start with specific groups of respondents, for instance with farms in the non-household sector (e.g.enterprises, cooperatives).In all cases, adequate testing of e-learning solutions as well as of CAWI, CATI and mail methods is critical prior to their adoption.

Map.
Status of censuses of agriculture (WCA 2020 round, May 2022) Note.The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of FAO concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers and boundaries.Dashed lines on the map represent approximate border lines for which there may not yet be full agreement.Final boundary between the Sudan and South Sudan has not yet been de ned.Dotted line represents approximately the Line of Control in Jammu and Kashmir agreed upon by India and Pakistan.The nal status of Jammu and Kashmir has not yet been agreed upon by the parties.Source: Castano (2022b).