FOOD SECURITY IN THE FACE OF COVID-19 Evidence from Africa Akuffo Amankwah and Sydney Gourlay As the pandemic started spreading to implementation varies across countries, sub-Saharan Africa, one concern has been as does the intensity of the pandemic and that of its possible impacts on food secu- the local restrictions (see Annex I for an rity, as the crisis has the potential to ex- illustration of survey timing and govern- acerbate an already fragile food security mental COVID-19 response). environment. Prior to the onset of the The analysis from the HFPS shows that COVID-19 pandemic, 55.6% of the sub-Sa- over 105 million adults are affected haran African population were moderately by moderate or severe food insecu- or severely food insecure.1 The health impli- rity across Uganda, Nigeria, Malawi, cations, movement restrictions, food supply Ethiopia, and Burkina Faso following disruptions, and other shocks brought on by the onset of the COVID pandemic, es- the pandemic may inflict increasing food se- timated using the Food Insecurity Expe- curity concerns across the region. rience Scale (FIES), the official methodol- Data from a series of high-frequency ogy for measuring SDG Indicator 2.1.2.3 phone surveys (HFPS) allows for the anal- In what follows, changes in food security ysis of food security challenges in the from pre-COVID periods and the preva- midst of the COVID-19 environment. The lence of food insecurity are explored, as HFPS data used here have been collect- well as the mechanisms behind food inse- ed primarily by national statistics offices2 curity in the post-pandemic world. in five SSA countries, with support from the World Bank’s Living Standards Mea- PREVALENCE OF FOOD INSE- surement Study (LSMS) and the Pover- CURITY: PAST AND PRESENT ty and Equity Global Practice. These five Food security has been a significant con- countries are part of the LSMS-Integrated cern for the region, even before the on- Surveys on Agriculture (LSMS-ISA) project slaught of challenges brought about by that fields longitudinal, multi-topic house- the COVID-19 pandemic. According to the hold surveys with a focus on agriculture. State of Food Security and Nutrition in the Thus, the households included in the HFPS World (2020), the prevalence of moderate were also interviewed as part of the LSMS- or severe food insecurity in the countries ISA panel survey conducted in these coun- reviewed here was as high as 82.2% of tries. A uniform methodology was adopted the population in the years preceding the in sampling, weighting, and implementing pandemic (Burkina Faso - 47.7%; Ethiopia the HFPS across the countries, making - 57.9%; Malawi – 82.2%; Nigeria - 44.1%; cross-country comparison feasible. While Uganda - 66.3%).4 the phone surveys began after the onset 3 The FIES questionnaire module used for SDG reporting uses a of the coronavirus pandemic, the timing of 12-month reference period. However, the LSMS-supported high- frequency phone surveys used a reference period of 30 days in order to assess food security and changes therein related to the COVID- 1 FAO, IFAD, UNICEF, WFP and WHO. 2020. The State of Food 19 pandemic period. The FIES questionnaire module is available at Security and Nutrition in the World 2020. Transforming food systems http://www.fao.org/in-action/voices-of-the-hungry/fies/en/ for affordable healthy diets. Rome, FAO. 4 FAO, IFAD, UNICEF, WFP and WHO. 2020. The State of Food 2 The Ethiopia HFPS was implemented by a private survey firm, not Security and Nutrition in the World 2020. Transforming food systems the national statistics office. for affordable healthy diets. Rome, FAO. 2 By employing the FIES methodology, The HFPS data also reveals a relationship we estimate the overall food insecuri- between food insecurity and well-being ty rates for both moderate and severe across all countries. Leveraging the con- food insecurity among the adult popla- sumption indices of the pre-COVID LSMS- tion after the onset of the pandemic.5,6 ISA surveys, we estimate the prevalence As illustrated in Figure 1, over 70% of of food insecurity by pre-COVID consump- adults in Nigeria and Malawi are im- tion quintile. Figure 2 illustrates this rela- pacted by moderate or severe food in- tionship, with households in the lower security, as well as 47% in Ethiopia, 42% end of the consumption distribution in Burkina Faso, and 43% in Uganda. presenting a higher rate of both mod- Over 30% of adults in Nigeria and Malawi erate and severe food insecurity, par- are plagued by severe food insecurity, as ticularly in Burkina Faso, Ethiopia, Malawi, well as 9% of Ugandan adults, 8% of Burki- and Uganda. nabe adults, and 13% of Ethiopian adults.7​ Food insecurity appears to affect rural households disproportionately vis-à-vis FIGURE 1. PREVALENCE OF MODERATE OR SEVERE urban households in Burkina Faso, Ethi- FOOD INSECURITY, BY COUNTRY opia and Malawi, with a greater share of the rural population experiencing mod- 80 erate or severe food insecurity (Figure 3). In Nigeria and Uganda, there is no signifi- 60 cant distinction between the prevalence of % of adults moderate or severe food insecurity across 40 urban and rural populations. The rate of severe food insecurity between urban and 20 rural populations is only statistically differ- ent in Malawi, with 33.4% of rural adults 0 Burkina Faso Ethiopia Malawi Nigeria Uganda and 23.4% of urban adults experiencing Moderate or Severe Severe severe food insecurity. What exactly does food insecurity look 5 Prevalence of moderate and severe food insecurity estimated like on a daily basis? The FIES questions, according to the FIES methodology, which employs item response theory. For details on the methodology, visit http://www.fao.org/ in-action/voices-of-the-hungry/en/. which include eight questions on behav- 6 Note that figures on FIES reflect the following survey rounds unless otherwise indicated: Burkina Faso – Round 2 (August); Uganda iors related to household food availabili- – Round 1 ( June 2020); Nigeria – Round 2 ( June 2020); Malawi – Round 1 (May/June 2020); Ethiopia – Round 3 ( June 2020). Estimates ty, allow us to assess the degree to which are made for the adult population rather than the population as a whole as relevant for SDG 2.1.2, in order to address inconsistencies households restricted food consumption. in the phrasing of FIES questions across countries. Malawi (Round 1), Nigeria (Round 2), and Uganda (Round1) phrased the FIES questions As illustrated in Figure 4, food consump- such that they asked specifically about adult household members. In subsequent survey rounds, these questions were revised to ask about any household member, regardless of age. Ethiopia and tion was a source of worry for the ma- Burkina Faso surveys ask about any household member. 7 Given the different reference population and reference period jority of households in all countries, of the HFPS FIES questions vis-à-vis the standard FIES and SDG reporting methodology, the estimates presented here are not with as many as 71% of households having comparable to the pre-COVID estimates reported in the State of Food Security and Nutrition in the World (2020). at least one member worrying they would 3 FIGURE 2. PREVALENCE OF FOOD INSECURITY, BY PRE-COVID CONSUMPTION QUINTILE 80 60 % of adults 40 20 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Burkina Faso Ethiopia Malawi Nigeria Uganda Moderate or Severe Severe not have enough food to eat (Burkina Faso at least one member going at least one – 69%; Ethiopia – 57%; Malawi – 71%; Nige- whole day without eating in Malawi and ria – 71%; Uganda – 58%). The majority of Nigeria. Eleven percent of households in households also had a member that was Ethiopia, 11% of households in Uganda, forced to skip at least one meal in Mala- and 20% of households in Burkina Faso wi and Nigeria, while more than 30% of also had a member that went at least one households had a member that skipped a whole day without eating. meal in Burkina Faso, Uganda, and Ethio- Food insecurity concerns pre-date the pia. Food consumption was severely lim- COVID-19 crisis, so how much of the food ited for a large share of households, with insecurity observed today can be attribut- approximately 36% of households having ed to the current crisis? Leveraging the FIGURE 3. PREVALENCE OF FOOD INSECURITY, BY RURAL/URBAN Prevalence of Moderate or Severe Prevalence of Severe Food Insecurity, Food Insecurity, by Rural/Urban by Rural/Urban 40 30 % of adults % of adults 20 10 0 Burkina Faso Ethiopia Malawi Nigeria Uganda Burkina Faso Ethiopia Malawi Nigeria Uganda Rural Urban Rural Urban 4 FIGURE 4. INCIDENCE OF HOUSEHOLDS WITH A LEAST ONE MEMBER WORRIED ABOUT FOOD, SKIPPING A MEAL, OR GOING A WHOLE DAY WITHOUT EATING, IN THE 30 DAYS PRECEDING THE INTERVIEW Burkina Faso Ethiopia Malawi Nigeria Uganda 0 20 40 60 80 % of households Worried Skipped Whole day panel nature of the HFPS data, and the food insecure).10 consistent FIES questionnaire implemen- Figure 5 depicts, at the household level, tation in Nigeria, we compare the food in- the movement from a food insecure to a security rates before COVID-19 (from the food secure status, and vice versa, for both 2018-19 GHS-Panel) and after the onset of moderate or severe and severe food inse- the pandemic (from the HFPS).8 Data from curity in Nigeria.11 Forty-three percent of the 2018-19 GHS-Panel, supported by the households that were not severely food LSMS-ISA, suggest that 48.5% of the Nigeri- insecure in 2018 were estimated to be an adult population suffered from moder- severely food insecure in June 2020, ate or severe food insecurity, while 14.0% representing a dramatic increase likely at- suffered from severe food insecurity.9 tributable at least in part to the COVID-19 The HFPS, implemented following the on- pandemic. The incidence of moderate or set of the COVID-19 pandemic, suggests severe food insecurity among the sample that the prevalence of food insecurity also increased significantly, with 71% of amongst the adult population in Nigeria households that were considered as food has increased to 75.5% (moderately or se- secure in July 2018 were moderately or se- verely food insecure) and 33.5% (severely verely food insecure in June 2020. 10 As a robustness check, we also limited the analysis sample to households that were interviewed in July 2018 to be compared with the estimates from HFPS interviews conducted in June 2020, both 8 This analysis is only possible for Nigeria given the comparability in the pre-harvest phase of the agricultural calendar (resulting of the questionnaires pre- and post-COVID. in a sample of 892 households). The 2018-19 data suggests that 46.8% (13.9%) of adults were moderately or severely (severely) food 9 Figures are reported from the post-planting visit of the GHS-Panel insecure. The post-COVID-19 outbreak data suggests a significant for better comparability with the HFPS interview timing. increase in the prevalence of food insecurity since July 2018, with 10 As a robustness check, we also limited the analysis sample to 75.1% of adults being moderately or severely food insecure and households that were interviewed in July 2018 to be compared with 32.6% severely food insecure. the estimates from HFPS interviews conducted in June 2020, both in the pre-harvest phase of the agricultural calendar (resulting 11 For the purposes of comparing food insecurity status over time, in a sample of 892 households). The 2018-19 data suggests that households are assigned to a food insecurity class based on the 46.8% (13.9%) of adults were moderately or severely (severely) food probability that adult members are food insecure. That is, if the adult insecure. The post-COVID-19 outbreak data suggests a significant members of a household have a probability greater than 50% then increase in the prevalence of food insecurity since July 2018, with they are moderately or severely food insecure, and the household is 75.1% of adults being moderately or severely food insecure and classified as such. 32.6% severely food insecure. 5 FIGURE 5. TRACKING FOOD INSECURITY OVER TIME FOR NIGERIA. No No No Yes Yes Yes 2018 June 2020 Moderately or Severely Food Insecure Severely Food Insecure The figures compare the food insecurity status of a subset of Nigerian households in 2018/2019 and June 2020 (graphic unweighted) Of the households that were not severe- UNPACKING FOOD INSECURITY ly food insecure in 2018 but became se- IN THE COVID-19 WORLD verely food insecure in June 2020, 35% re- Of the many shocks endured by house- sided in urban areas, significantly higher holds since the onset of the pandemic, than the share of urban households that increase in the price of major food items were not severely food insecure in both consumed by the household was one 2018 and June 2020, suggesting that ur- of the most prevalent. In Malawi, 66% of ban households in Nigeria are dispropor- all households reported an increase in tionately affected by severe food insecu- the price of key consumption goods (July rity following the onset of the pandemic. 2020), while the same was true for 90% Conversely, those in the wealthiest con- of Nigerian households (July 2020) and sumption quintile were disproportionately 53% of households in Burkina Faso (Au- food secure following the pandemic, when gust 2020). Increases in the market price looking both at severe and moderate food of consumption goods will harm the food insecurity. Patterns also emerge across security of households, particularly those Nigeria’s geography, with the South West without the ability to produce their own seeing more households shift to severe food or in the lower end of the consump- food insecurity than remaining not severe- tion distribution. While there seem to be ly insecure, and the North Central zone similar distribution of households report- seeing more households shift to moder- ing experiencing shocks due to increase in ate or severe food insecurity than remain- the price of food items consumed across ing secure.12. rural-urban divide in Nigeria and Burkina Faso, the impact appears to be more se- vere among rural households in Malawi 12 More detailed findings on the characteristics of households shifting from a food secure to insecure status have been omitted for (Figure 6). brevity, but are available upon request. 6 FIGURE 6. PREVALENCE OF FOOD PRICE SHOCKS, BY COUNTRY (LEFT) AND URBAN/RURAL STATUS (RIGHT) Prevalence of Food Price Shocks, Prevalence of Food Price Shocks, by Country by Rural/Urban 90 80 % of households % of households 70 60 50 Malawi Nigeria Burkina Faso Malawi Nigeria Burkina Faso Urban Rural FIGURE 7. PREVALENCE OF FOOD PRICE SHOCKS TO MAJOR CONSUMPTION GOODS (% OF HOUSE- HOLDS), BY CONSUMPTION QUINTILE 100 80 % of households 60 40 Malawi Nigeria Burkina Faso 1 2 3 4 5 Analysis of the incidence of price shocks Food consumption, though a basic across the consumption distribution re- need, is one of the levers used by many veals a statistically similar incidence across households to cope with shocks of all consumption quintiles, particularly for Ni- types. In July 2020, 66% of Nigerian house- geria and Burkina Faso (Figure 7). In Mala- holds reduced food consumption as a cop- wi, households in the wealthiest quin- ing mechanism in response to a variety of tile experienced a lower incidence of shocks. The same is true for 9% of house- food price shocks, suggesting that poor- holds in Malawi and 16% of households in er households suffered disproportionate- Uganda. Nine percent of households lim- ly from food price shocks. ited food consumption in response to in- 7 FIGURE 8. PERCENT OF HOUSEHOLDS REDUCING FOOD CONSUMPTION TO COPE WITH SHOCKS OVER TIME, BY CONSUMPTION QUINTILE. 80 60 % of hosueholds 40 20 0 Ethiopia Malawi Nigeria Uganda Burkina Faso 1 2 3 4 5 FIGURE 9. PERCENT OF HOUSEHOLDS REDUCING FOOD CONSUMPTION TO COPE WITH SHOCKS OVER TIME, BY CONSUMPTION QUINTILE. 80 60 % of households 40 20 0 1 2 3 4 5 1 2 3 4 5 Ethiopia Nigeria April/May June/July come loss in Ethiopia (June 2020).13 as the pandemic progresses, more house- The HFPS provides evidence that the re- holds in the lower quintiles are restricting striction of food consumption as a means of food consumption due to shocks. Figure coping with shocks is a behavior observed 9 presents the share of households that in households across the consumption dis- reduced food consumption in two consec- tribution. As illustrated in Figure 8, there is utive months for Nigeria and Ethiopia.14 a fairly consistent implementation of food In both countries, the prevalence of this restriction across pre-COVID-19 wealth food reducing behavior increased in the quintiles. There is evidence, however, that second month of interview, and it in- 13 In Ethiopia, coping mechanism questions were presented to 14 Ethiopia and Nigeria were the only two countries to include respondents in the context of their response to income loss rather the relevant data in more than one interview, as of the date of than to shocks in general. publication 8 creased more for households on the low end of the consumption distribution. The change was extremely pronounced in Ni- geria, where there is a dramatic increase in the restriction of food consumption of the households in the poorest two quintiles as the pandemic state extend- ed. This may suggest that the longer the COVID-19 environment persists, the more vulnerable households will be impacted, and potentially to a disproportionate degree. 9 ANNEX I. COUNTRY-LEVEL Malawi COVID RESPONSE & HFPS INTERVIEW TIMING 100 The figures below illustrate the timing 80 of each HFPS survey round against the 60 COVID-19 Government Response Strin- 40 gency Index.15 Only HFPS survey rounds that are analyzed in this brief are included. 20 In all countries, subsequent survey rounds 0 have been or will be collected. The survey January 1 2020 March 1 2020 May 1 2020 July 1 2020 Sep 1 2020 round dates presented below are trimmed HFPS COVID-19 Government Response Stringency Index (5%) to eliminate outliers. Burkina Faso Nigeria 100 100 80 80 60 60 40 40 20 20 0 0 January 1 March 1 May 1 July 1 Sep 1 January 1 March 1 May 1 July 1 Sep 1 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 HFPS COVID-19 Government Response Stringency Index HFPS COVID-19 Government Response Stringency Index Ethiopia Uganda 100 100 80 80 60 60 40 40 20 20 0 0 January 1 March 1 May 1 July 1 Sep 1 January 1 March 1 May 1 July 1 Sep 1 2020 2020 2020 2020 2020 2020 2020 2020 2020 2020 HFPS COVID-19 Government Response Stringency Index HFPS COVID-19 Government Response Stringency Index 15 Thomas Hale, Sam Webster, Anna Petherick, Toby Phillips, and Beatriz Kira (2020). Oxford COVID-19 Government Response Tracker. Last updated Nov. 5, 2020. 10 FOOD SECURITY IN THE FACE OF COVID-19 EVIDENCE FROM AFRICA JANUARY 2021 WWW.WORLDBANK.ORG/LSMS