The 50x2030 Initiative and production of SDG 2 indicators: Country challenges and experiences

2.1Agricultural Integrated Surveys as a source for SDG 2

The 50x2030 survey system, similar to the AGRIS methodology, was designed following the modular approach: an annual core questionnaire focused on crop, livestock, aquaculture, fishery, and forestry production, and a set of specialized rotating questionnaires covering topics such as costs and farm income, labor and productivity, gender decision-making in agriculture, and production practices and environmental aspects of farming. Table 1 presents an overview of the integrated questionnaires and the coverage of SDG indicators.

The sequence of modules determines the frequency and order of the production of the Indicators, as can be deduced from Table 1 and Fig. 1. Since the standard system has a three-year cycle, the complete set of the three Indicators are produced every three years. SDG Indicators 2.3.1 and 2.3.2 are produced in the first year of the cycle, while Indicator 2.4.1 is produced after the first two years of each cycle, according to distribution of its sub-Indicators shown in Table 2. Computation of the SDG Indicators 2.3.1, 2.3.2, and 2.4.1 presupposes a dataset of agricultural holdings with comprehensive coverage of the agricultural activities of the country. This is crucial for determining the thresholds for the definition of the range of smallholder units in the case of SDG 2.3.1 and 2.3.2 and for producing reliable estimates of labor productivity, average income, and the country agricultural area in the context of SDG 2.4.1.

The integrated survey approach involves the development and use of a Master Sampling Frame (MSF) and a sampling design that ensures completeness of a country’s agricultural activity, targeting all agricultural production units (i.e., agricultural households and corporate producers) and covering all crop and livestock activities. The use of multiple frames is also recommended, usually consisting of an area frame and/or list frame for the household sector and a list frame for the non-household sector or special farms. Section 3.2 includes details on the sampling frames used by the analyzed countries for the surveys utilized to produce the Indicators. Using a single statistical survey system ensures consistency in data collection, in terms of concepts, reference period, and information completeness.

Another central aspect of the integrated survey approach is the expansion of the survey scope. In general, agricultural surveys are focused on strictly agricultural data, such as cultivation area, yield, and crop production, and animal numbers. Taking advantage of the modular approach, the integrated system goes far beyond traditional agricultural statistics, collecting data related to multiple aspects associated with agriculture and allowing for the inclusion of aquaculture, fishing and forestry activities of the agricultural holdings. It seeks to reconcile more lengthy questionnaires with data quality, minimising respondent burden and ensuring the production of detailed data, as the SDGs require.

2.2SDG 2 computation requirements

Indicators 2.3.1, 2.3.2 and 2.4.1 are the outputs of computational processes that follow refined methodologies, which also include common elements, such as the need to identify the small-scale food producers (SDG 2.3.1; SDG 2.3.2), calculate their volume of production and their agricultural land area (SDG 2.3.1; SDG 2.3.2; 2.4.1) and production costs (SDG 2.3.2; 2.4.1) [9, 10].

SDG Indicator 2.3.1 measures the agricultural labor productivity, i.e., the agricultural volume of production per labor input, whereas the SDG Indicator 2.3.2. measures the average income of small-scale food producers, where income refers to:

The formulas of the above-mentioned Indicators are the following:

where

The agricultural activities considered are crop land farming, fishing capture, aquaculture, livestock farming and forestry activities, whereas the agricultural land is that under operation, including the area cultivated by the agricultural holdings (regardless of land ownership) with temporary and permanent crops, aquaculture land and temporary pasture, land rented in and fallow land. It excludes land that is rented out, forestland and land that had been abandoned prior to the reference period. Moreover, – all values in local currency units need to be converted in Purchasing Power Parity Dollars (PPP $) for countries comparison purposes.

The SDG indicator 2.4.1 for the year t measures the proportion of agricultural area under productive and sustainable agriculture in the year t, which is expressed by the following formula:

where the denominator includes the area under temporary and permanent crops, meadows, pastures, fallow land and other land used for agriculture, plus forest and other wooded land and areas dedicated to aquaculture. The focus of this Indicator is those holdings whose primary activities are crop cultivation and/or livestock grazing. Holdings that have other activities as their primary focus are excluded from the calculation. The value of Indicator 2.4.1. is calculated by aggregating the results of each of the 11 sub-indicators, that measure sustainability in different areas. They are: 1. Farm output value per hectare; 2. Net Farm Income; 3. Risk mitigation mechanisms; 4. Prevalence of soil degradation; 5. Variation in water availability; 6. Management of fertilizers; 7. Management of pesticides; 8. Use of agrobiodiversity-supportive practices; 9. Wage rate in agriculture; 10. Food Insecurity Experience Scale (FIES); and 11. Secure tenure rights to land.

Table 3 shows the components required to calculate each Indicator and the level of accessibility of that information through the agricultural surveys, as evaluated from different countries’ experiences.

2.3SDG 2 sampling requirements

Sampling procedures should be carefully considered in the design of integrated surveys to ensure the reliable measurement of SDG 2 Indicators. Accordingly, the 50x2030 Initiative proposes an integrated sampling design to fulfil the measurement objectives of the Integrated Agricultural and Rural Survey Program [11]. Specific sampling requirements for SDG 2 Indicators are highlighted below.

The sampling frame should adequately cover both the agricultural sector and the household sector (in order to satisfy the requirements of some Indicators). The use of a multiple frame would be necessary to fulfil the following requirements:

The sample size should be large enough to ensure reliable estimations of the target SDG Indicators. The sample size calculation procedures based on the measure of statistical dispersion (coefficient of variation or standard error) of a key variable of interest in the population is preferable. Target variables correlated to the various SDG 2 Indicators should be considered. For integrated surveys, agricultural land or value of production and household income or consumption is adequate. The 50x2030 Initiative discusses the calculation of an integrated sample size suitable for both agriculture and household-related indicators [11].

Stratification is an important operation for improving the precision of estimates. Since strata are Indicator-specific, the choice of stratification criteria suitable for the various SDG Indicators is important, so as to avoid having too many strata, which can be problematic. In multistage sampling, stratification of primary sampling units (PSUs) should be carefully performed, not only for improved estimates but for suitable coverage of sub-populations for reporting purposes. Criteria based on the structure of households in PSUs e.g., by practice of agriculture, could be helpful [11]. For the specific case of SDG 2.4.1, additional criteria suitable for the various sub-indicators could be considered, including the type of activity (crops, livestock and mixed holdings), irrigation (irrigated/non-irrigated holdings) or soil fertility-based criteria [12].

SDG 2 Indicators are expected to be reported at the national level. To reduce respondent burden and related measurement errors, the questionnaires of the specific modules designed for the SDG can be administered only to a subsample. Subsampling would also be cost-effective for objective measurements, e.g., plot land measurement using GPS or yield estimation with crop cutting operations.

Moreover, following the principle of “leaving no one behind,” SDG indicators are supposed to be disaggregated, where relevant, by income, sex, age, race, ethnicity, migratory status, disability and geographic location, or other characteristics. The sampling design should allow, as much as possible, the observation of sampling units for every sub-population for which disaggregated data must be produced. Sampling procedures for disaggregated SDG Indicators are discussed in [13].

For a sustainable measurement of SDG Indicators, sampling procedures over time should consider the frequency of SDG reporting, sample obsolescence over time, the cost of updating samples over time and the requirement of estimating changes in the Indicators for performance assessment. The 50x2030 Initiative recommends either panel or partial rotation sampling over time as cost-effective sampling approaches for computing SDG 2 Indicators [14].

Figure 2.

AGRIS modular architecture (recommended module flow).

3.1Integrated survey system implementation: Thematic priorities and SDG2

Since 2018, countries assisted by the AGRISurvey programme have adopted the integrated survey approach, progressively adjusting their existing survey to ensure comprehensive coverage of the national agricultural activity and expanding the survey scope by incorporating the rotating modules.

Most countries prioritise the economic aspects with specific interests in farm income, cost of production, or even to inform the national accounts. In the sequence, priority is given to environmental aspects; frequency varies but the three-year cycle is most common. It is important to mention that commonly key labour aspects, in particular labour input, were incorporated in the core module. In this case, data for computation of SDG Indicator 2.3.1 is obtained every year (e.g., Cambodia, Ecuador, Georgia).

Table 6 synthesizes country implementation, describing the order and frequency of the rotating modules, and indicating plans for the production of SDG 2 Indicators from the survey rounds in subsequent years. The combination of rotating modules is also observed, even though it represents longer interviews, with excessive respondent burden and some loss of data quality.

Besides the institutional, financial, and operational aspects, the implementation of the integrated approach mainly involves elaborating or revising the sampling frame and sampling design and developing data collection instruments. The latter activity, in principle, refers to the customization of standard questionnaires according to the country’s characteristics. In practice, the general configuration of the survey system could differ substantially from country to country. It is important to distinguish between cases where implementation is undertaken from scratch and the reformulation of a survey system that is already in place. In the first case, usually, the standard methodology is adopted. In the second, characteristics of the previous, ongoing survey system tend to be preserved, and adaptations developed (e.g., Ecuador, Georgia, Senegal, Uganda). The mode of implementation varies according to the type of institution in charge (e.g., national statistical office or ministry of agriculture) and basic survey objectives. The interest in early crop estimates and/or area measurement or crop-cutting determines the adoption of several data collection waves during the year, corresponding to each agricultural season (e.g., Senegal and Uganda) and season cycle (post-planting and post-harvesting time); while the interest in providing data for national accounts could claim for a quarterly survey implementation (e.g., Georgia). Moreover, the interest mainly in agricultural official statistics and cost-effectiveness would tend towards the one-visit approach (e.g., Cambodia and Ecuador). The different configurations have an impact on the complexity involved in computing the Indicators.

3.2The coverage issue

Limitations regarding the coverage of agricultural activities are not rare in agricultural survey systems, particularly among developing countries. Besides a lack of financial or operational resources, the following can be mentioned:

There are normally very few corporation-like agricultural holdings compared to family farms. However, even in countries where the sector has very low participation, it could be relevant for a particular agricultural activity/commodity or in terms of land possession or value of production. Administrative or private sector data are normally limited in scope. This data could complement or serve as a benchmark during the data validation process, but it hardly comprises the information required regarding statistical units or variables.

As mentioned before, a multiple-frame survey design is necessary. Traditional sampling schemas, such as two-stage sampling or area frame sampling based on segments, have been the preferred approaches to ensure the general coverage of the household sector. For instance, Cambodia, Senegal, Uganda, Nepal and Georgia use a list frame for the household sector of the agricultural surveys where they extract the Indicators. In particular, Uganda and Georgia benefit from a rich master sampling frame that provides samples for all the agricultural surveys and also, in the case of Uganda, for households surveys. Ecuador instead makes use of an area frame for the household sector. All the countries analyzed in this paper associate to the frame for the household sector a second frame for the non-household sector and special units. Building this type of frame is always challenging. Difficulties include the definition of a legal corporation itself, considering local legislation and formal record-keeping practices in the country; the identification, access and quality of cadastres; and integration of records from competent administrative or private-sector data (e.g., Cambodia, Senegal and Uganda). Due to the relevance of the problem, special guidelines on the topic were developed in the context of the AGRISurvey programme [28].

The simulation summarized in Table 7, using the Georgia Survey of Agricultural Holdings 2020, demonstrates a possible impact on estimates when the non-household sector is neglected. The smallholder population that should be the target of a specific policy is underestimated, as well as labor productivity and average income of the policy target group. Senegal and Uganda (AAS 2019) discarded the non-household sector in the computation of both SDG Indicators 2.3.1. and 2.3.2. Tables 8 and 9 provide an overview of the coverage and thematic scope of analyzed countries’ surveys, with reference to the data required for SDG Indicators 2.3.1, 2.3.2 and 2.4.1.

Comprehensive coverage of a country’s agricultural activities is crucial for adequate computation of SDG Indicators 2.3,1 2.3.2, and 2.4.1. The misclassification of corporate farms and other legal entities involved in agricultural production shifts the determination of the threshold to define the smallholder group, tending to underestimate the target population, as in the case of SDG Indicators 2.3.1 and 2.3.2. The lack of coverage of all types of producers and agricultural activities harms the estimates of all Indicators mentioned. Ensuring adequate coverage is the first challenge faced by countries. The adoption of an integrated survey system helps to overcome the issue.

It was observed that in most countries, certain regions are intentionally left out of the survey due to marginal relevance, like large cities (all countries), the absence of agriculture activities or remoteness and/or hard accessibility (e.g., Ecuador) or even due to inescapable political or military conflicts or obstacles (e.g., Georgia). Punctual coverage issues due to short-term inaccessibility such as flooding events were also observed (e.g., Indonesia pilot). During the last two years, the Covid-19 pandemic has affected coverage but mainly resulted in delays in data collection or adaptation of data collection modes to phone-based interviews using tablets, Computer-assisted telephone interviewing (CATI) or Computer-assisted web interviewing (CAWI).

Beyond the comprehensiveness of the target population and sampling design, attention should also be paid to the questionnaire content. In some situations, the complete absence or insufficient coverage concerning relevant agricultural activities in the data collection instruments was observed. For example, forestry and fisheries information was not collected in the case of the Ecuador ESPAC, Georgia SAH, Nepal pilot 2019, Uganda AAS 2019, Senegal EAA 2017/2018. The prices of forestry products were not collected in the Cambodia Inter-Censal Agriculture Survey 2019 (CIAS 2019), making it impossible to calculate the value of production generated by forestry activities which was, consequently, excluded from both SDG 2.3.1 and 2.3.2 computation. In some other cases, questions related to minor activities or less relevant commodities were avoided in favor of justified reduction of the respondent burden and costs, with gains in additional survey efficiency (e.g., Cambodia). For instance, the value of production and costs related to crop activities took into consideration only the 30 main crops in Cambodia since detailed information on minor crops was not collected in CIAS 2019.

3.3Required questions and questionnaire size and flow

As mentioned, the integrated approach proposed by the 50x2030 Initiative combines comprehensive coverage of agricultural activities and the most relevant themes related to the agricultural and rural sectors in the same survey system. It is conceived as a way of guaranteeing not only comprehensiveness in terms of coverage and thematic enlargement but also of ensuring concept standardization, result coherence, resource sharing, field- and deskwork optimization, and, in particular, cost-effectiveness.

The integrated approach also contributes to greater efficiency in data collection. By combining different modules applied to the same sample and operation, it is possible to obtain a larger amount of coherent information at a lower cost and in an efficient manner. It also expands the analytical possibilities, through crosschecking the data collected and the production of more complex Indicators, such as the farm survey-based SDG 2 Indicators. The FAO Working Paper on Mainstreaming SDG Indicator 2.4.1 maps the correspondence between the proposed SDG 2.4.1 Indicator stand-alone questionnaire and the 50x2030 tools [23].

Typically, an integrated survey system combines multiple objectives, resulting in richer but lengthier questionnaires. So, careful consideration should be taken regarding the trade-off between thematic coverage, respondent burden, and data quantity and quality. In practice, implementation of the system involves customization of the standard data collection instruments. Customization involves adaptation of the standard instruments and recommendations according to specifics, such as the country’s agricultural characteristics and policy priorities, respondents’ profiles, and the country’s technical and resource capacities.

The requirements for the numerator of SDG 2.3.1 correspond to the traditional production information collected annually in agricultural surveys and are part of the core module. Moreover, all assisted countries have expressed interest in maintaining the collection of key labour data in the core module, centered on annual labour input, which is the denominator of the Indicator. The core module, as well, provides the necessary data for the determination of the thresholds that define the smallholder sub-population. As a result, the SDG 2.3.1 Indicator on labour productivity can be produced every year for countries adopting the expanded core module.

This is not the case for SDG 2.3.2, since farm income requires data related to costs of production and holding expenses, which could be collected periodically using the economic modules. The collection of this kind of data is very demanding, involving information related to several items. Nevertheless, countries have decided to include the cost and expenses data with high frequency, such as every year, every other year, or every three years as the standard proposal suggests.

With 11 Sub-indicators, SDG 2.4.1 on agricultural sustainability involves the collection of several types of non-traditional data, including aspects related to environmental issues and the social dimension. The mode of measurement of the 11 SDG 2.4.1 Sub-indicators is very precisely specified and incorporated in the 50x2030 data collection instruments and some country surveys. According to the theme, the required questions are allocated to different sections of the questionnaire in order to ensure better logic and coherence with the flow of the questionnaire.

It is important to mention that there are several synergies between the three SDGs. Careful measurement of the value of production is fundamental; it is required to compute both SDG 2.3.1, SDG 2.3.2 and the first Sub-indicator of SDG 2.4.1. The costs related to agricultural activities are also the main components of SDG 2.3.2 and the SDG 2.4.1 second Sub-indicator if the simplified option is not applied. The criteria for identifying the smallholders relevant to SDG Indicators 2.3.1 and 2.3.2 are the same, meaning that time and resources can be saved if both data for both indicators are collected and computed together.

The inclusion of the required questions/sections could significantly increase the length of the questionnaire. As a result, the computation of SDG 2.4.1 in some countries was only partial. The questions for the 2.4.1 Sub-indicator 10 on Food Insecurity Experience Scale (FIES) and for some more demanding environmental Sub-indicators were either dropped (e.g., Nepal pilot 2019) or planned to be incorporated into different survey editions (e.g., Cambodia). The FIES indicator was dropped in the case of the Georgia SAH 2021, due to lack of relevance.

To reduce respondent burden and survey cost, both the 50x2030 Initiative and the AGRIS recommendations suggest applying the rotating modules to a sub-sample. In this case, the focus is the production of the SDGs Indicators at the national level, with sub-national level results obtained only for less desegregated domains. This recommendation was followed by some assisted countries, however, most opted to apply the rotating modules to the entire sample.

Some issues have been identified relating to certain aspects of the surveys and questionnaires. Since production data is often collected on a declaration basis at the holding level, the main issue is the use of non-standard units by respondents, making it difficult to produce estimates on standard units without appropriate conversion factors. This is an important source of measurement error for this kind of data. To ensure quality production data, a detailed list of conversion factors that allow converting crop harvests from local to metric units should be developed, even if it requires putting in place a dedicated survey to update the coefficients. To help countries facing this issue, the 50x2030 Initiative proposes developing a technical note describing such an ad hoc survey [29]. Some countries (e.g., Senegal) have developed and integrated a specific questionnaire into the main survey to collect additional data on the conversion factors.

Another issue is the lack of producers’ price data in the questionnaire to enable valuation of the production. In some cases, the customized questionnaires did not include commodities and/or input prices. Minimal attention is devoted to forestry activities, particularly the collection of price information on forestry products, making it difficult to estimate the value of production (e.g., Cambodia, CIAS 2019). In this case, administrative data may be used to calculate the required values. Recall issues are a common source of bias concerning: the selling price of agricultural products; agricultural production; production lost, destroyed or given to service or input providers; and quantities of agricultural inputs. Some input costs (e.g., hybrid and local seeds) were impossible to estimate (e.g., Nepal pilot 2019). The costs of several agricultural inputs and labor inputs were not collected and most of them were obtained through administrative sources.

Some proxies and assumptions were used to harmonize measurement units, such as the hypothesis that the annual labor input of a permanent worker corresponds to 250 man-days, since the unit of measurement of temporary worker labor and exchange labor available in the questionnaire was man-days (Nepal pilot 2018).

For the calculation of some Sub-indicators of SDG Indicator 2.4.1, the simplified option is often chosen, including in the questionnaire proxy questions. For example, in order to compute the second Sub-indicator of SDG 2.4.1, the direct question on profitability in the last three years is used instead of more detailed questions on a holding’s revenues over the same period.

For Indicator 2.4.1, the measurement of the agricultural area is crucial. Questions relating to land area and use are asked in a very detailed manner, collecting area in the following categories: area under temporary crops; fallow; meadows and pastures; permanent crops; other types of area used for agriculture; forest and other wooded lands; and area dedicated to aquaculture. Measurement errors can affect the precision of the agricultural land area reported by farmers through declaration, considering also the fact that it comprises several components. One of the strengths of some of the survey programs studied is the objective measurement of land area, which contributes to reliable yield estimates and more accurate production values. However, production data based on farmer estimates are subject to: (i) recall bias, and (ii) distortions in the conversion from local to standard units. This important measurement error affects crop, livestock, aquaculture and forestry products, and it is a common source of errors in agricultural surveys in developing countries. Similarly, the cost of inputs, including labor input (time), are affected by recall mistakes. The time worked on the holding is particularly difficult to collect due to the seasonality, irregularity and informality of farm labor.

The content of each Agricultural Integrated Survey questionnaire is quite complex and creates respondent burden during enumeration. The median amount of time is nearly one hour, when focused on the core questionnaire alone, and can reach two hours or more when rotating modules are included. This situation is known to jeopardize the quality of responses. The duration of the interview, however, is also associated with the complexity of the farm activity, in particular, it can be related to the number of parcels and plots when questions are not asked at the holding level but at the plot level. One strategy utilised has been to limit the number of crop items for which detailed information is captured, neglecting the information regarding less relevant ones. A multiple-visit strategy was also considered for some types of respondents.

Table 10 shows the interview times for statistical operations in different countries, in terms of the number of visits and the rotating modules used.

3.4Data collection challenges

As mentioned, some information required for the computation of the farm survey-based SDG 2 Indicators is standard, such as crop and agricultural area; livestock numbers; agricultural production; and use of irrigation, fertilizer and other inputs. However, most of the required information corresponds to aspects more recently incorporated into agricultural surveys, linked to the large scope of the emerging data needs. Moreover, many of the new themes of interest correspond to phenomena that are hard to measure, representing an additional challenge to producing consistent and reliable estimates of the Indicators. Information that is difficult to capture includes the following:

3.5Indicator computation and challenges due to missing components

In this section, challenges encountered when performing SDG computations due to missing components or difficulties in methodology interpretation are discussed. Information on how the challenges were overcome is provided, including (i) deviations from the strict Indicator methodology eventually adopted; (ii) proxies used for the calculation due to missing questionnaire items; and (iii) the use of administrative data or other sources to complement data requirements. The challenges and solutions described could refer to one or more country experiences.

The main deviation from the strict indicator methodology is in the identification of the smallholder group required for the computation of SDG 2.3.1 and 2.3.2. As indicated above, in the absence of non-household sector holdings, proper identification of the smallholder group, which requires the inclusion of the entire population of holdings, is problematic. Given that farm size distributions are typically highly skewed (i.e. there is a large number of small farmers and a low number of large farmers) and have “long tails” (a very low number of very large farmers, which significantly impacts the aggregate land size), the approach, based on 40% of the cumulated value, will inevitably lead to identifying a very large proportion of farmers as small food producers. This proportion will be even higher if agricultural surveys are used, as opposed to household surveys, as the former also includes commercial farms. That is one reason for using agricultural surveys to determine the set of small-scale food producers and to calculate SDG Indicators 2.3.1 and 2.3.2. The use of household surveys as an alternative should be considered only when non-household sector agricultural production is considered absolutely negligible. While household surveys may indeed contain most of the data required to determine the set of small-scale food producers and compute the SDG Indicators, they only cover the household sector and therefore farms that tend to be smaller, generate lower incomes and possess a lower number of animals than in the non-household sector. Determining the set of small-scale food producers from these surveys, as opposed to sector-wide agricultural surveys, would generally lead to lower thresholds and likely distortions in the results, as farm size (physical and economic) is correlated to average farm incomes and productivity. For some countries, the problem could be overcome only in the subsequent editions of the surveys, incorporating a list of special (non-household) farms in the frame and sample of the survey.

Another deviation from the methodology of Indicators 2.3.1, 2.3.2 and 2.4.1 is the exclusion of the forestry and fishery sectors in the computation of the Indicators. Few attempts have been made to estimate the missing information, i.e., the value of aquaculture, fishery and forestry production or the prices of their products and consequently the relative values of production. Therefore, these sectors are sometimes excluded from the calculation of the Indicators. Note, however, that in general, they represent marginal activities of the agricultural holdings. Therefore, their absence is not considered a concern. In any case, it could have a consequence for evaluation of the “on-farm diversification” in the context of the SDG 2.4.1 Sub-indicator 3 – Risk mitigation mechanisms.

3.6Publication of Indicators: clearance and dissemination

The 50x2030 Initiative promotes best practices in partner countries that aim to make statistics accessible to the public in formats that ensure greater readability, usability, interoperability, and findability of the data. In line with countries’ data dissemination strategies (if available), targeted support is provided to improve dissemination policies and programs, helping ensure that dissemination programs become part of national institutional processes.

To the extent possible, the Initiative promotes the dissemination of timely statistics, recommending that the data should be disseminated within six to 12 months after data collection is completed and target data users/producers and donor communities at the national level.

At the United Nations-level, the release of the SDG Indicators is guided by the official principles for submission and publication of the SDG indicators, which are contained in the IAEG-SDG Guidelines on Global Reporting [35]. The main principle is that countries should report to the relevant custodian agency or publish their data on their own national reporting platform. However, there is no official procedure for this, because the specifics may vary from Indicator to Indicator. It is often the case that countries submit the underlying data and FAO calculates the final figure based on that data. Usually, it is good practice to allow for both scenarios, i.e., a country can proceed to the final calculation itself or it can leave this to FAO. By contrast, it is not good practice to allow countries to simply send the final figure to the custodian agency without any underlying data (although exceptions could be made on a case-by-case basis), due to the need for quality assurance procedures performed by the custodian agency.

On the other hand, at the country level, the dissemination of SDG Indicators 2.3.1, 2.3.2 and 2.4.1 is country-specific, according to each country’s National SDG programs, release calendars and priorities. In most countries, the governments prepare SDG priority plans. The implementation and monitoring of the plans are different from country to country Some of them develop a detailed program, which determines the national institutions responsible for monitoring the SDGs, the data providers, the reference years and the targets adjusted to the national context (e.g. Georgia) [36]. Other countries establish coordination groups that gather members of the relevant Ministries and NSOs (e.g., in Cambodia the coordination groups are the Data for Development Group, led by NIS and the SDG Secretariat 2019, located in the Office of the Prime Minister) or incorporate the SDG goals into the broader national development policy (e.g., Senegal integrated the SDG goals into its national development policy 2019–2023). However, in general, the countries do not adopt a strict national SDG publication plan and the calculated SDG is not included in the survey publications (Georgia is an exception). One of the objectives in the dissemination program of AGRIS and 50x2030 Initiative is closely linked to assistance in the calculation of the selected SDG Indicators, namely 2.3.1, 2.3.2, 2.4.1, and 5.a.1, and their proper dissemination, either through the country websites (e.g. the NIS website in Cambodia, Geostat website in Georgia) or through FAO platforms. Table 6 shows the progress achieved, or planned, in the years during the period of assistance provided in the countries. The procedures for the publication of SDG Indicators need to be further developed and well communicated to the countries.