ROUND #1 COVID-19 IMPACT MONITORING JUNE 2020 Publication Date BACKGROUND Although the first case of COVID-19 in Uganda was confirmed on the 22nd of March, Government of Uganda had under- UGANDA taken several actions starting on the 18th of March, including travel restrictions, a 14-day quarantine for all international arrivals, and cancellation of all international conferences and public gatherings, including, but not limited to, religious ser- vices, weddings and concerts. On the 30th of March, the President declared a nationwide curfew from 7 pm to 6:30 am; banned public transportation; and instituted strict regulations for the movement of government and private vehicles. Con- tainment measures, regional instability and broader trade uncertainty are expected to negatively affect economic activities, growth and incomes. Poverty is expected to increase as a result of the impact of COVID-19 on trade and services, while lower internal and external demand for agricultural products will deteriorate rural incomes. In June 2020, the Uganda Bureau of Statistics (UBOS), with the support from the World Bank, has officially launched the High-Frequency Phone Survey on COVID-19 to track the impacts of the pandemic on a monthly basis for a period of 12 months. The survey aimed to recontact the entire sample of households that had been interviewed during the Uganda Na- tional Panel Survey (UNPS) 2019/20 round and that had phone numbers for at least one household member or a reference individual. Of 2,421 households that were attempted to be interviewed, 2,259 were successfully interviewed, representing 93 percent of the initial target sample. Gender distribution of the respondents was close to parity. This brief presents find- ings from the first round of the survey that was conducted during the period of June 3-20, 2020. KNOWLEDGE, BEHAVIOR AND CONCERNS RELATED TO COVID-19 TRANSMISSION Knowledge of COVID-19 symptoms is still far from uni- versal. The most frequent COVID-19 symptoms such as dry cough, fever and shortness of breath were in fact not men- tioned by all respondents. Eighty three percent of respond- ents reported dry cough to be a COVID-19 symptom, and there were no significant differences in reporting by the lev- el of respondent’s education. On the other hand, while fever was mentioned by 67% of respondents, the awareness of this symptom was significantly lower among those that never attended school (48%). Only 36% of respondents named shortness of breath as a COVID-19 symptom and almost nobody mentioned loss of smell or taste (4%). Figure 1. Knowledge of selected COVID-19 symptoms Awareness of preventive measures is quite high, but with gatherings (98%), wearing a mask (95%), social distancing some variation across rural/urban areas and pre-COVID (91%) and avoiding touching the face (87%). Key preventive -19 consumption quintiles based on UNPS 2019/20. The measures such as handwashing and social distancing were respondents were well-informed about the important pre- universally known. ventive measures such as handwashing (100%), avoiding a) at the national level b) by the level of respondent’s education Figure 2. Knowledge of selected measures to reduce the risk of contracting coronavirus (% of respondents) 1 COVID-19 IMPACT MONITORING However, the poorest 20% of the respondents as well re- educated are less aware of some preventive measures such as spondents that are living in rural areas and those that are less using sanitizers and gloves and avoiding touching the face. False beliefs regarding COVID-19 coexist with accurate consumption of alcohol provides immunity to the disease; 22 knowledge about symptoms and preventive measures % are of the opinion that Africans are immune to COVID-19; and are correlated with the level of respondent’s educa- and 16% do not think that children can be affected by COVID tion. Large groups of respondents reported false beliefs re- -19. Almost all false beliefs are strongly correlated with lower garding COVID-19 which may discourage hygiene and social levels of education. False beliefs may contribute to lower distancing. 44% of the respondents believe that coronavirus adoption of preventive measures despite very high shares of will not survive in warm weather; 38% believe that local respondents (more than 95%) who claimed to have altered herbs can help treating COVID-19 patients; 26% think that their behavior in favor of safe practices. a) at the national level b) by the level of respondent’s education Figure 3. Share of respondents who either believe in false statements regarding COVID-19 Absolute majority of respondents consider coronavirus as threat to their household’s finances compared to rural re- a threat both to health and financial status of their spondents (84%), the level of concern in fact does not vary households. 76% of respondents worry about themselves or by pre-COVID-19 household per capita consumption quin- their immediate family members becoming seriously ill due to tile. 89 % of respondents living in the poorest 20 % of house- the coronavirus. Even more respondents (86%) perceive a holds perceive a threat to their household’s finances due to threat to their household's finances from the coronavirus. the coronavirus, while the comparable estimate is 88 % While urban respondents (90%) are more likely to feel a among those living in the richest 20 % of households. ACCESS TO BASIC NEEDS Soap and water Access to water is not an issue but having enough soap is a problem among the poorest households, mainly due to economic reasons. Hand hygiene remains one of the most effective actions to reduce the spread of COVID-19. Each respondent was asked whether his/her household had enough soap to wash hands and whether the household had access to water during last 7 days. Less than 1% of house- holds had issues accessing water, while almost 18% of house- holds struggled having enough soap to wash hands. This share is largest in rural areas (20%) and among the poorest households (30%). Absolute majority of those who did not have enough soap point to economic reasons: could not af- ford it (67%), no cash to buy (13%) and high prices (8%). Figure 4. Share of households without enough soap to wash hands 2 COVID-19 IMPACT MONITORING Main staple and non-staple food items Access to main staple and non-staple food items con- sumed with staples remains relatively high. Each respond- ent was asked to (i) specify the main staple food item, and separately, the main non-staple food item consumed with the staple item (referred to as “sauce” in the interview) for his/ her household, and (ii) report on the household’s need and ability to buy these items during the week preceding the sur- vey. Most households needed to buy the main staple, but the ratio is lower in rural (72%) than urban areas (85%). Among those that reported the need to buy their main staple food item, 16% could not do so. Similarly, of those that needed to buy their main non-staple food item, 19% could not do so. Otherwise, there are no clear differences in access to main staple and non-staple food items across the pre-COVID-19 household per capita consumption quintiles. Figure 5. Share of households who needed to buy food items, and share of households who were not able to do so, conditional on need Economic reasons restrict access to main staple and non poorest 20 percent of households and in rural areas are -staple food items. The main reasons for not being able to more likely to report lack of cash as the main reason, while buy these items were (i) increase in price and (ii) lack of cash those living in wealthier households and in urban areas are (and inability to use a credit card). The respondents living in more likely to be affected by increase in prices. Health services Rural households experienced more issues accessing of households needed medicine during this reference period. medicine compared to their urban counterparts. Re- Among those that needed medicine, 33% could not access it. spondents were asked about their households’ need for The comparable estimate was 36% among rural households medicine during the week preceding the survey. About 80% versus 26% among urban households. Figure 6. Share of households who needed medicine and Figure 7. Reasons of not being able to access medical medical treatment, and share of those who were not able treatment when needed, by urban/rural to do so, conditional on need Rural residents experienced more issues accessing medi- during this reference period. Rural households reported cal treatment - mainly due to lack of money and lack of higher need for medical treatment than urban ones. Among transportation. Respondents were asked about their those who needed medical treatment, 19% could not access households’ need for medical treatment since March 20. it. The comparable estimate was 21% among rural house- About 57 percent of households needed medical treatment holds versus 15% among urban households. 3 COVID-19 IMPACT MONITORING For both rural and urban residents, lack of money was the percent of urban households could not access medical treat- key reason for the unmet need. Lack of transportation was ment due to the travel restrictions imposed by the authori- also very important for rural households, while almost 10 ties. Education School closures widen pre-existing inequalities in access to schooling. All schools were closed in Uganda on the 20th of March. Before the pandemic, distribution of households with any child age 3-18 enrolled in educational institutions were relatively equal across residence and pre-COVID-19 consumption quintiles. Overall, 92% of households with at least one child in the age group of 3-18 had at least one child enrolled in school prior to the closure of schools. After the closures, the share of households with any child attending any remote learning activity stands at 59% and is distributed very unequally. For example, it ranges from 44% among the poorest quintile to 74% among the richest quintile. There is also a statistically significant gap between rural and urban Figure 8. Share of households with a child (3-18) in school areas. prior to closures vs. share of households with any child par- ticipating in remote learning activities after closures Learning activities for children following the school clo- sures exhibit differences across rural and urban areas to differences in electricity access and ownership of TVs, and across the pre-COVID-19 household per capita con- radios, and phones. Children from rural households and sumption distribution. Children engage in numerous differ- households from first and second poorest consumption per ent learning activities, but far the most wide-reaching activi- capita quintiles are more likely to listen to education pro- ties include using reading materials provided by the govern- grams on radio and use reading materials provided by gov- ment, listening to radio, watching educational programs on ernment. Children from urban areas and wealthiest fourth TV and completing assignment provided by teacher. The and fifth consumption per capita quintiles are more likely to types of distance learning activities differ across urban and watch education TV programs and use mobile learning appli- rural areas and across consumption per capita probably due cations. Figure 9. Share of households with students participating in specific educational activities during the school closures, by rural/urban (conditional on having at least one school-age child attending learning activities). Food security The survey asked eight food security questions which allow rural and urban areas and by pre-COVID-19 household per the construction of the Food Insecurity Experience Scale. capita consumption quintiles. Frequency Survey on the Im- We break down the selected food security indicators across pacts of COVID-19 on Households is a collaboration of … 4 COVID-19 IMPACT MONITORING Overall, in the last 30 days preceding the interview, 8% households experienced severe food insecurity while 42% experienced moderate or severe food insecurity. There were no statistically significant differences across urban and rural areas. However, households from the poorest con- sumption quintiles, in particular, the bottom 40 percent, are more likely to experience moderate or severe food insecurity. Figure 10. Share of households with severe versus mod- erate or severe food insecurity, by rural/urban and pre- COVID-19 household per capita consumption quintile. EMPLOYMENT AND LIVELIHOODS Job losses, and changes in working conditions among wage employees COVID-19 had a considerable impact on the working status of the respondents to the survey. The vast majority of the respondents were still working the week before the interview (70% on average), more than half of the non- working respondents stopped working after the restrictions put in place in response to the pandemic. Respondents in urban areas and those living in households in the top 40% of the pre-COVID-19 per capita consumption distribution suf- fered from job interruptions the most. More than 17 per cent of respondents in Central and Eastern Uganda stopped working after March 20, when the Government of Uganda closed schools and public offices. At the national-level, among 90 percent of respondents that stopped working cited COVID-19-related reasons for job interruptions. Figure 11. Status of employment last week, by rural/ urban/regional residence and pre-COVID-19 household per capita consumption quintile. Respondents who stopped working in the post-March 20 period were overwhelmingly employed in sectors that entail the most personal interactions. Overall, more than one third of these individuals were working in the commerce sector (i.e. buying and selling), while 24% were working in the service sector (i.e. personal services). There are notable differences in the sectoral composition of job interruptions across rural and urban areas. 40% of urban respondents who lost their jobs were working in the commerce sector, while 30% were working in the service sector. In rural areas, about one third of respondents who lost their jobs were working in agriculture, while 28% were working in the commerce sector and about 17% were em- ployed in the service sector. Figure 12. Work stoppages, by industry of main job. 5 COVID-19 IMPACT MONITORING Service, transport and commerce are the sectors hit the most by the COVID-19 restrictions, having lost the high- est share of workers. 43% of respondents in the service sector (i.e. personal services) were no longer working dur- ing the last week preceding the survey interview. The com- parable estimates were for those in the transport and com- merce (i.e. buying and selling) sectors were 39% and 34%, respectively. Agriculture was the least impacted sector. Indeed, 93% of respondents in this sector kept working despite the anti-COVID19 measures. Figure 13. Status of employment last week, by industry of main job. Income Changes Since the COVID-19 outbreak, 87% of households have ily business suffered income losses (less or no earnings) sub- reported reduced income (or no earnings) from at least sequent to the COVID-19 outbreak. A contraction in trans- one of their sources of livelihood. Figure 14 provides the fers from-family within the country was observed among shares of households receiving income from specific sources 83% of households that received this type of income in the over the last 12 months. Figure 15 shows the reported last 12 months. The comparable incidence of income loss or change in income since March 20 by income source, condi- no earnings was 65% among those that have received wage tional on having received income from that source over the employment income in the last 12 months and 60% among last 12 months. 90% of households involved in non-farm fam- households involved in farming. Figure 14. Household income sources in the last 12 Figure 15. Changes in income since March 20, 2020. months. 6 COVID-19 IMPACT MONITORING The incidence of non-farm business ownership was high- have reported losses subsequent to the COVID-19 out- est among households in the top 40 percent of the pre- break. 97% of these losses are due to a reason potentially COVID-19 household per capita consumption distribu- related to the COVID-19. Although, only 6% of non-farm tion. Most of these businesses are in the commerce sector, family businesses are in the transport industry, all have been which was the hardest hit by the pandemic. As shown in Ta- reported to have experienced a contraction in revenue. ble 1 below, 94% of households participating in the sector Table 1. Family business - Revenues by enterprise Current sales revenue (late March/ April), Potentially re- % of HHs with compared to February 2020* lated to COVID family business -19 Less Same Higher Buying & Selling 68 94 5 2 97 Personal Services 10 91 8 1 97 Mining 9 10 90 0 100 Transport 6 100 0 0 98 AGRICULTURE Crop farming On the whole, the share of farming households increased from 72% in 2019 to 78% in 2020. The increase was 10 per- centage points among those in Central Uganda, and 8 percentage points among households in Western Uganda as shown in Figure 16. The increase in the incidence of farming was most pronounced among those in the top 20% of the pre-COVID-19 house- hold per capita consumption distribution. Figure 16: Share of households engaging in crop farming during the first seasons of 2019 versus 2020, by region and pre-COVID-19 household per capita consumption quintile. 23% of the households that were engaged in farming in On the other hand,38% of the households that have changed 2020 reported that the COVID 19 pandemic had influ- crop planting activities as a result of COVID-19 reduced enced their crop cultivation decisions. 38% of these crop area under cultivation, particularly in Eastern Uganda. households increased crop area under cultivation and 17% While 11% reduced the diversity of crops cultivated on the increased the diversity of crops cultivated on the farm. In- farm, especially in Central Uganda. Further, 6% of house- crease in area planted was the most frequent change in plant- holds abandoned farming due to COVID 19 and this practice ing activities in in Northern Uganda and in the poorest con- was most common in Western Uganda. Finally, the 5% of sumption quintile. The surge in the diversity of crops culti- households delayed planting due to COVID-19 and particu- vated on the farm is the preferred strategy for households in larly in Eastern Uganda. The main reported reasons for Western Uganda and for those that are in the top 40% of changing crop planting activities were being advised to stay the pre-COVID-19 household per capita consumption distri- home (51%), movement restrictions (42%), lack of availability bution. of labor (17%) and lack of other input availability (6%). 7 COVID-19 IMPACT MONITORING Figure 17: Share of households undertaking specific changes to crop planting activities due to COVID-19, by region and pre- COVID-19 household per capita consump- tion quintile (Q1-Q5). Livestock production Only 8% of the livestock-keeping households reported reported probably because the use of Artificial Insemination that COVID-19 has affected their livestock production (AI) is very rare in Uganda. Changing the feed ration was the activities. Among the livestock keeping households that most frequent effect of COVID-19 on livestock production were affected by COVID-19, 52% changed feed ration due to in Central Uganda and Western Uganda while failure to ac- costs, 40% could not access veterinary services, 30% could cess to veterinary services was the most reported effect as not vaccinate their animals, 25% could not deworm their in Eastern Uganda. Inability to sell animals was most fre- animals,15% had to change animal watering regime, and 15% quently reported among the household in the lowest con- could not sell their animals as shown in Figure 18. The post- sumption quintile (Q1), while changing feed ratio was associ- ponement of Artificial insemination service (AIs) was not ated with households in the higher consumption quintiles. Figure 18. Effects of COVID-19 on live- stock production activities, by region and pre-COVID-19 household per capita con- sumption quintile (Q1-Q5). Sale of the Agricultural Outputs COVID-19 has affected agricultural households' ability ern Uganda (31%) and Northern Uganda (30%). The regions to sell their outputs due to closure of weekly and with the higher share of households that needed to and monthly markets as well as travel restrictions. Overall, were able to sell their products (namely Central and West- 44% of households needed to sell farm produce. Among ern Uganda) were also the ones with the higher share of these households, 41% could not see their produce - corre- households that needed to but were unable to sell their sponding to 18% of all farming households, irrespective of products. Also, the need to sell and being able to sell was their need to sell agricultural outputs. Western Uganda had reported most by households in higher consumption quintile highest percentage of farmers that needed to sell the agricul- as shown in Figure 19. tural produce (57%) followed by Central Uganda (55%), East- 8 COVID-19 IMPACT MONITORING Figure 19: Effects of COVID 19 on sale agricultural produce by region and pre-COVID-19 household per capita con- sumption quintile (Q1-Q5). SAFETY NET Though not shown, the incidence of household cash transfer receipts from social assistance pro- grams is less than 1%. At the national-level, 9% of households have received food aid since March 20. In urban areas, the incidence of food aid receipt is 17%, while the comparable statistic is 6% among rural households. Breaking down the incidence of food aid receipt by pre-COVID-19 household per capita consumption quintiles reveals that food transfers are not reaching the poorest and are in fact disproportionately targeted towards the rich- est. The incidence of food aid receipt is 5% in the poorest quintile, while the comparable statistic is 16% in the richest quintile. Though not reported, the national-level differences across consumption quintiles are driven specifically by what is happening in urban areas. These findings call for further re- search into and discussion regarding the process of targeting food transfers, particularly in urban areas and as a function of employment status and expo- Figure 20. Incidence of Food Aid Receipt, by Rural/Urban and pre- sure to shocks, among other factors. COVID-19 household per capita consumption quintile (Q1-Q5). Figure 21 reports incidence of household exposure to selected shocks since March 20. The most common shock is by far the increase in the prices of food items consumed. At the national-level, 29% of households are reported to have been exposed to increases in food pric- es. The second-most common shock is non-farm business failure, underlining again the adverse effects of COVID-19 on the informal economy and livelihoods. At the national- level, 14% of households reported to have been exposed to non-farm business failure, with no differences across rural and urban areas. Finally, the third-most common shock is the fall in prices of outputs produced by house- holds. In this case, the incidence of household exposure is 11 percent at the national-level, 13% among rural house- holds and 8% among urban households. Figure 21. Incidence of selected shocks, by rural/urban. 9 COVID-19 IMPACT MONITORING Figure 22 reports incidence of selected coping strategies tion quintiles. The extent of reliance on savings is 51 percent among the sub-sample of households that were exposed to among households in the richest quintile versus 37% among at least one shock since March 20. At the national-level, 23 those in the poorest quintile. The second-most common percent of households did nothing in response to the shock coping strategy is reduction in food consumption. At the – an estimate that was the highest for the poorest first pre- national-level, the incidence of reduction in food consump- COVID-19 household per capita consumption quintile. The tion among households exposed to shocks is 28 percent, most common coping strategy was reliance on savings, with with the comparable estimates in the top and bottom 20 the national-level incidence of 43% disguising important dif- percent of the pre-COVID-19 consumption distribution be- ferences by pre-COVID-19 household per capita consump- ing 19 percent and 26 percent, respectively. Figure 22. Distribution of Coping Strategies Among Households Exposed to Shocks, by pre-COVID-19 household con- sumption quintile. Data Notes: Uganda High-Frequency Phone Survey (HFPS) on COVID-19 is implemented by the Uganda Bureau of Statistics (UBOS) during the period of June 2020-May 2021. The survey is part of a World Bank-supported global effort to support countries in their data collection efforts to monitor the impacts of COVID-19. The financing for data collection and technical assistance in support of the Uganda HFPS COVID-19 is provided by the United States Agency for International Development (USAID) and the World Bank. The technical assistance to the survey is provided by a World Bank team composed of staff from the Development Data Group - Living Standards Measurement Study (LSMS) program and the Poverty and Equity Global Practice. In Round 1, 2,421 households that had been previously interviewed during the 2019/20 round of the Uganda National Panel Survey (UNPS) were contacted, and 2,257 households were successfully inter- viewed, with the goal of re-interviewing them in the subsequent monthly HFPS COVID-19 rounds. The pre-COVID-19 UNPS data are nationally-representative and the survey weights were calcu- lated for the HFPS sample (i) to counteract selection bias associated with not being able to call UNPS households without phone numbers, and (ii) to mitigate against non-response bias associat- ed with not being able to interview all target UNPS households with phone numbers. For further details on the data, visit http://www.worldbank.org/lsms-covid19. 10