{"doc_desc":{"prod_date":"2023-10-30T04:00:00.000Z","producers":[{"name":"Bo Pieter Johannes Andree","abbr":"B.P.J. Andree","affiliation":"World Bank","role":"Lead WFSO"}],"idno":"WLD_2023_WFSO_v01_M","version_statement":{"version":"v1.1","version_date":"2024-11-30T23:00:00.000Z","version_resp":"Bo Pieter Johannes Andree, World Bank, Development Economics Data Group"},"title":"World Food Security Outlook"},"study_desc":{"title_statement":{"title":"World Food Security Outlook","idno":"WLD_2023_WFSO_v01_M","alternate_title":"WFSO","identifiers":[{"type":"DOI","identifier":"https:\/\/doi.org\/10.48529\/ev5a-ke69"}]},"series_statement":{"series_name":"Other Household Health Survey [hh\/hea]","series_info":"The World Food Security Outlook (WFSO) database, provided by the World Bank Development Economics Data Group and the Agriculture Global Practice, tracks and analyzes global food security. It includes historical, preliminary, and forecast data on severe food insecurity worldwide. "},"version_statement":{"version":"October 2025","version_date":"2025-11","version_resp":"Bo Pieter Johannes Andree","version_notes":"This version is based on the IMF\u2019s World Economic Outlook of October 2025 and the latest data from the World Development Indicators, FAO\u2019s national statistics (FAO-STAT) on food insecurity and global food prices (FAO-FPI), the Bank\u2019s RTFP data on food prices in data-poor regions, and FEWS NET assessments. The data do generally not reflect the impact of events up to the cut-off date as there is varying delay with which official data is published.\n\nWFSO data are not directly comparable across different versions. Each year, with the publication of the annual SOFI report, FAO re-estimates entire data series, revising or occasionally deleting historical data. Consequently, WFSO estimates are also re-evaluated based on the updated data and may change over time. When entire country data series are deleted, WFSO estimates may still reflect earlier SOFI data. This is due to the WFSO estimation process, which employs a Markov Chain Monte Carlo algorithm that initializes with values from the previous WFSO, following the updating methodology introduced by Andr\u00e9e and Pape (2023). As a result, if FAO removes country data, the algorithm may still begin with values close to previous SOFI estimates, allowing past data to influence new WFSO calculations. However, this influence diminishes over time.\n\nFAO does not maintain a public archive of older data versions. Nevertheless, historical versions of WFSO are preserved for transparency and research purposes. In rare cases, where FAO country series are highly inconsistent with credible food security reports, the statistics may be replaced with estimates derived from IPC-compatible data, which are clearly marked in the WFSO notes.\n\nUsers are encouraged to access data directly from FAO STAT to ensure they are using the most up-to-date official figures. It is worth noting that WDI may sometimes lag behind FAO STAT in reflecting official data updates."},"study_info":{"abstract":"Key components of the WFSO database cover the [prevalence of severe food insecurity](https:\/\/data.worldbank.org\/indicator\/SN.ITK.SVFI.ZS), including estimates for countries lacking official data, population sizes of the severely food insecure, required safety net financing, and corresponding estimates expressed on the Integrated Phase Classification ([IPC](https:\/\/www.ipcinfo.org\/)) scale. Data is presented in a user-friendly format.\n\nWFSO data primarily relies on hunger and malnutrition data from the State of Food Security and Nutrition in the World (SOFI) report, led by the Food and agriculture Organization (FAO) in collaboration with multiple UN agencies. WFSO complements SOFI data by providing estimates for unreported countries. Historical estimates are produced with a [machine learning model](http:\/\/hdl.handle.net\/10986\/38139) leveraging [World Development Indicators](https:\/\/databank.worldbank.org\/source\/world-development-indicators) (WDI) for global coverage. This model has been extended to express outputs on the IPC scale by converting estimates using a nonlinear beta regression estimated on a normalized range, and distributionally adjusted using a [smooth threshold transformation](http:\/\/documents.worldbank.org\/curated\/en\/304451600783424495).\n\nFinancing needs for safety nets are calculated similarly to past approaches by the International Development Association (IDA) to assess food insecurity response needs ([IDA (2020)](https:\/\/documents1.worldbank.org\/curated\/en\/775981606955884100\/pdf\/Responding-to-the-Emerging-Food-Security-Crisis.pdf) and [IDA (2021)](https:\/\/documents1.worldbank.org\/curated\/en\/252271636587686210\/pdf\/IDA19-Mid-Term-Review-of-the-Crisis-Response-Window-Early-Response-Financing.pdf)). Preliminary estimates and projections rely on the same model and incorporate International Monetary Fund (IMF)'s [World Economic Outlook](https:\/\/www.imf.org\/en\/Publications\/WEO) (WEO) growth and inflation forecasts. WEO data reflects the IMF's expert analysis from various sources, including government agencies, central banks, and international organizations.\n\nMinor gaps in WDI data inflation data are replaced with unofficial WEO estimates. Minor inflation data gaps not covered by both, are replaced with unofficial inflation estimates from the World Bank's [Real Time Food Prices](https:\/\/microdata.worldbank.org\/index.php\/catalog\/4509) (RTFP) data.\n\nThe WFSO is updated three times a year, coinciding with IMF's WEO and SOFI releases. It provides food security projections that align with economic forecasts, aiding policymakers in integrating food security into economic planning.\n\nThe WFSO database serves various purposes, aiding World Bank economists and researchers in economic analysis, policy recommendations, and the assessment of global financing needs to address food insecurity.\n\nAdditionally, the WFSO enhances transparency in global food security data by tracking regional and global figures and breaking them down by individual countries. Historical estimates support research and long-term trend assessments, especially in the context of relating outlooks to past food security crises.","data_kind":"Process-produced data [pro]","analysis_unit":"Country","keywords":[{"keyword":"Food Security"},{"keyword":"Macro-economic outlook"},{"keyword":"Inflation"},{"keyword":"Fragility, Conflict and Violence"},{"keyword":"Economic Outlook"},{"keyword":"SDG2"},{"keyword":"Hunger"},{"keyword":"Malnutrition"},{"keyword":"Food Crisis"},{"keyword":"FIES"},{"keyword":"Food Insecurity"},{"keyword":"Humanitarian Aid"}],"geog_coverage":"World","geog_coverage_notes":"191 countries and territories mutually included by the World Bank's WDI and IMF's WEO databases. The country coverage is based on mutual inclusion in both the World Bank World Development Indicators database and the International Monetary Fund\u2019s World Economic Outlook database. Some countries and territories may not be covered. Every attempt is made to provide comprehensive coverage. To produce complete historical predictions, missing data in the WDI are completed with unofficial data from the WEO and the World Bank's RTFP data when inflation data is not available in either database. Final gaps in the WDI and WEO are interpolated using a Kernel-based pattern-matching algorithm. See background documentation for equations.","geog_unit":"Country","coll_dates":[{"start":"1999","end":"2030","cycle":"annual"}],"quality_statement":{"other_quality_statement":"The WFSO data and metadata provided are \"as is\" and \"as available,\" and every effort is made to ensure their timeliness, accuracy, and completeness. When errors are discovered, they can be reported to the principal investigator and will be corrected as appropriate and feasible. "},"nation":[{"name":"World","abbreviation":"WLD"}],"topics":[{"topic":"Q18 - Agricultural Policy; Food Policy","vocab":"JEL"},{"topic":"F35 - Foreign Aid","vocab":"JEL"},{"topic":"I32 - Measurement and Analysis of Poverty","vocab":"JEL"},{"topic":"O13 - Economic Development: Agriculture; Natural Resources; Energy; Environment; Other Primary Products","vocab":"JEL"},{"topic":"Q01 - Sustainable Development","vocab":"JEL"},{"topic":"I14 - Health and Inequality","vocab":"JEL"}]},"authoring_entity":[{"name":"Bo Pieter Johannes Andree","affiliation":"World Bank"}],"production_statement":{"funding_agencies":[{"name":"Federal Ministry for Economic Cooperation and Development as part of the World Bank\u2019s Food Systems 2030 Multi-Donor Trust Fund ","abbr":"BMZ","grant":" TF073570 and  TF0C0728","role":"Maintenance of model outlook"}]},"oth_id":[{"name":"Food and Agriculture Organization (FAO)","affiliation":"United Nations","role":"Source of primary data"}],"method":{"method_notes":"The original data is constructed based on an assessment that is conducted using data collected with the Food Insecurity Experience Scale or a compatible experience-based food security measurement questionnaire (such as the HFSSM). The probability to be food insecure is estimated using the one-parameter logistic Item Response Theory model (the Rasch model) and thresholds for classification are made cross country comparable by calibrating the metrics obtained in each country against the FIES global reference scale, maintained by FAO. The threshold to classify \"severe\" food insecurity corresponds to the severity associated with the item \"having not eaten for an entire day\" on the global FIES scale. It is an indicator of lack of food access. The indicator is calculated as an average over 3 years (eg. data for 2015 is the average of 2014-2016 data). These data are then complemented by 3-year averages of population-weighted sub-national Integrated Phase Classification (IPC) compatible data produced by FEWS NET.  The IPC-compatible data is scaled to the prevalence rates using a linear fixed effects regression in countries where both data are available, and linear regression where only the IPC-compatible data is available. The resulting data is then modeled against select World Development Indicators using a Cubist regression validated and calibrated against the original prevalence rates. The Cubist regression model is maintained by the World Bank and used to predict unobserved entries, predict historical values before the original data series start, and predict future values based on WDI data complemented with forward-looking data from the IMF\u2019s World Economic Outlook. The result is combined with population data from the WDI and WEO to produce estimates of the number of severely food insecure. The number of severely food insecure are combined with data from the cost of a minimum caloric diet from SOFI (2019) adjusted for global food price inflation using the FAO Food Price Index. Future prices are extrapolated using a state-space model at the lower limit, point and upper limits of a 90% confidence range.","data_collection":{"time_method":"Panel survey [pan_svy]","instru_development":"Model is based on Andree, B.P.J. Machine Learning Guided Outlook of Global Food Insecurity Consistent with Macroeconomic Forecasts. Policy Research working paper; no. WPS 10202 Washington, D.C. :World Bank Group. [http:\/\/hdl.handle.net\/10986\/38139](http:\/\/hdl.handle.net\/10986\/38139)\n\nScale up to 191 countries was implemented in Gatti, R., Lederman, D., Islam, A.M., Bennett, F.R., Andree, B.P.J., Assem, H., Lotfi, R., Mousa, M.E. (2023). Altered Destinies: The Long-Term Effects of Rising Prices and Food Insecurity in the Middle East and North Africa. MENA Economic Update; April 2023. Washington, DC. World Bank. [http:\/\/hdl.handle.net\/10986\/39559](http:\/\/hdl.handle.net\/10986\/39559)\n\nThe financing needs are calculated following IDA (2020) Responding to the Emerging Food Security Crisis, and IDA (2021) Mid Term Review of the Crisis Response Window Early Response Financing. \n\nIPC-compatible estimates are produced by estimating a betaregression with areal IPC phase percentages (Andree et al, 2020) on the dependent side and prevalance rate and financing needs expressed as a share of GDP as predictors. The regression predictions are rescaled using a distributional transformation optimized using cross-validation (Andree et al, 2020). See:\n\nAndree, B.P.J., Chamorro, A., Kraay, A., Spencer, P.G., Wang, D. (2020). Predicting Food Crises. Policy Research Working Papers. World Bank. http:\/\/documents.worldbank.org\/curated\/en\/304451600783424495"}},"data_access":{"dataset_use":{"disclaimer":"For details on the terms and conditions for usage of the WFSO database, please refer to the license details. "}},"study_notes":"Development Relevance: Food insecurity at moderate levels of severity is typically associated with the inability to regularly eat healthy, balanced diets. As such, high prevalence of food insecurity at moderate levels can be considered a predictor of various forms of diet-related health conditions in the population, associated with micronutrient deficiency and unbalanced diets. Severe levels of food insecurity, on the other hand, imply a high probability of reduced food intake and therefore can lead to more severe forms of undernutrition, including hunger. FAO has identified the FIES as the tool with the greatest potential for becoming a global standard capable of providing comparable information on food insecurity experience across countries and population groups to track progress on reducing food insecurity and hunger."},"tags":[{"tag":"DOI"},{"tag":"test"},{"tag":"WFSO"}],"schematype":"survey"}