Monitoring Social and Economic Impacts of COVID-19 on Refugees in Uganda: Results from the High-Frequency Phone Survey- Third Round World Bank Poverty and Equity Practice May 18, 2021 Contents Acknowledgments......................................................................................................................................... 6 Key messages and findings ........................................................................................................................... 7 A. Background ......................................................................................................................................... 11 B. Economic activities.............................................................................................................................. 12 a. Labor market and farm activities .................................................................................................... 12 b. Changes in household income since the lockdown ........................................................................ 13 c. Impact on poverty and food security .............................................................................................. 14 d. Social assistance .............................................................................................................................. 15 C. Knowledge and behavior .................................................................................................................... 16 D. Access to goods and basic needs ........................................................................................................ 18 E. Coping strategies to socioeconomic shocks ....................................................................................... 20 F. Mental health, concerns and interactions with hosts ........................................................................ 21 G. Bibliography ........................................................................................................................................ 24 H. Annex .................................................................................................................................................. 25 a. Sampling.......................................................................................................................................... 25 b. Profile of respondents and households .......................................................................................... 26 2 List of figures Figure 1. Employment status of refugees (in last 7 days) across regions before the lockdown and in rounds 1-3, (% of all respondents) .......................................................................................................................... 12 Figure 2. Key reasons for work stoppages among the same refugee respondents across regions in rounds 1 and 3, (% of the respondents who stopped working).............................................................................. 12 Figure 3. Ownership of family business among refugee households in rounds 1-3, (% of households) .... 13 Figure 4. Changes in business revenues for rounds 2 and 3 (% of households with family business) ....... 13 Figure 5. Households with higher or the same level of key income sources in rounds 2 and 3 compared to the averages before the outbreak (% of refugee households) ................................................................... 13 Figure 6. Households with higher or the same level of key income as before the outbreak in rounds 2 and 3 across regions (% of refugee households) ............................................................................................... 13 Figure 7. Poverty rates before COVID-19 and across rounds using SWIFT methodology (% of refugee population).................................................................................................................................................. 14 Figure 8. Food security during last 30 days in rounds 1-3, (% of refugee households) .............................. 14 Figure 9. Food security during last 30 days across regions in rounds 1-3, (% of refugee households) ...... 15 Figure 10. Types of social assistance in rounds 1-3 (% of refugee households) ......................................... 15 Figure 11. Types of social assistance across regions in rounds 1-3, (% of refugee households) ................ 15 Figure 12: Prevalence of safe practices in rounds 1 (since March), rounds 2 and 3 (last week), (% of respondents) ............................................................................................................................................... 16 Figure 13: Prevalence of safe practices across regions in rounds 1 (since March), rounds 2 and 3 (last week), (% of respondents) ...................................................................................................................................... 16 Figure 14: Frequency of mask wearing when in public, in rounds 2 and 3 (%) and number of known people who wear masks most or all of the time when in public, in round 3 ......................................................... 16 Figure 15: Attendance of religious or social gatherings during last 7 days in round 3, (% of respondents) .................................................................................................................................................................... 16 Figure 16: Acceptance to be vaccinated if approved vaccine was available right now at no cost in rounds 2 and 3, (% of refugee respondents) ............................................................................................................. 17 Figure 17: Household with any member ever received a laboratory diagnosis for COVID-19 and respondents who would be willing to be tested at no cost in round 3, (%) ............................................... 17 Figure 18: No access to medicine during last month when needed by regions in round 3 (% of refugee households)................................................................................................................................................. 18 Figure 19: No access to medical treatment when needed by regions and across rounds 1-3 (% of refugee households)................................................................................................................................................. 18 Figure 20: Inability to buy main staple food when needed by regions and head of household gender in rounds 1-3, (% of refugee households)....................................................................................................... 19 Figure 21: Share of households unable to cover basic needs and access services among refugees and Ugandans (% of households) ...................................................................................................................... 19 Figure 22: Number of shocks in rounds 1-3, (% of refugee households).................................................... 20 Figure 23: Households experienced one shock only, (% of refugee households) ...................................... 20 Figure 24: Type of shocks in, (% of refugee households) ........................................................................... 20 Figure 25: Selected strategies to cope with socioeconomic shocks in rounds 1-3, (% of refugee households) .................................................................................................................................................................... 21 Figure 26: Respondents with major depression, (% of respondents) ......................................................... 21 Figure 27: How depression affects work, housework or getting along with other people, (% of refugee respondents who experience at least one depression symptom in two weeks) ....................................... 22 Figure 28: Perceived substantial threat from coronavirus pandemic to household’s finances, (% of refugee households)................................................................................................................................................. 22 3 Figure 29: Respondents who believed having had COVID-19 and knew someone who had it, cumulative from rounds 2 and 3, (% of respondents) ................................................................................................... 23 Figure 30: Respondents who did not feel safe, (% of respondents) ........................................................... 23 List of tables Table 1: Distribution of observations and population across strata in rounds 1, 2 and 3 .......................... 25 Table 2. Characteristics of respondents and households in the second round .......................................... 26 List of boxes Box 1: Selected findings from comparing refugees and Ugandans .............................................................. 9 4 © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. The material in this work is subject to copyright. Because the World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. 5 Acknowledgments This project is possible due to collaboration between the World Bank, Uganda Bureau of Statistics (UBOS) and United Nations High Commissioner for Refugees (UNHCR) supported by the Government of Uganda (GoU), particularly the Office of the Prime Minister (OPM). Phone survey data was collected by Laterite. The project was led by Aziz Atamanov and Nobuo Yoshida from the World Bank and Theresa Beltramo and Peter Waita from the UNHCR. The core team also included Charles Alemi (UNHCR), John Ilukor (World Bank), Laura Abril Rios Rivera (World Bank), Ibrahima Sarr (UNHCR), Ally Hamud Said (UNHCR), and Kazusa Yoshimura (World Bank). The report was prepared under the supervision of Antony Thompson (World Bank, Country Manager for Uganda) and Pierella Paci (World Bank, Practice Manager). Implementation of the survey was guided by Joel Boutroue (UNHCR Representative, Uganda) and Margaret Atieno (UNHCR Assistant Representative Protection, Uganda) from UNHCR. The authors are grateful to Stephen Baryahirwa, James Muwonge and Vincent Fred Ssennono from the Uganda Bureau of Statistics for collaboration. We are also thankful for excellent comments from Lilian Achieng, Zewditu Banteyehun Haile, Bo Hurkmans, Cara Ann Myers, Benjamin Christopher Reese, Shinya Takamatsu. Xiaomeng Chen and Dixita Gupta provided support in imputing consumption and estimating poverty among refugees. Kexin Zhang produced weights for the survey. Martin Buchara provided excellent administrative assistance. Finally, we would also like to thank all respondents of the survey. The State and Peacebuilding Fund (SPF) is a global fund administered by the World Bank to finance critical development operations and analysis in situations of fragility, conflict, and violence. The SPF is kindly supported by: Australia, Denmark, France, Germany, The Netherlands, Norway, Sweden, Switzerland, The United Kingdom, as well as IBRD. 6 Key messages and findings Economic impact: ✓ Employment rates among refugee respondents declined from 43 percent in October/November 2020 to 32 percent in February/March 2021. This widened the gap further from the pre-lockdown employment rate of 56 percent—by 24 percentage points. The decline in employment rates was driven by the West Nile region, where most work stoppages were related to seasonal changes. ✓ Ownership of family businesses declined from 37 percent before the lockdown in March 2020 to 23 percent in February/March 2021. Ownership of family business declined in the third round vis- à-vis second round in the West Nile region and remained stagnant in the South West and West Nile regions. This was accompanied by a slightly growing share of households with a declining or a complete loss of business revenues. ✓ Household income from the key sources did not deteriorate and slightly improved in the West Nile and South West regions with more households reporting income levels to be the same or higher than the pre-March 2020 levels. ✓ While the most severe forms of food insecurity declined slightly in February 2021 compared to October/November 2020, mainly in the South West and to a lesser extent in the West Nile region, refugees continued to face very high levels of food insecurity. ✓ The estimated poverty rates remained around 50 percent in rounds 2 and 3 and they did not return to the pre-COVID-19 level of 44 percent. Social assistance: ✓ The incidence of social assistance declined across rounds, with a gradual increase in the share of households that did not receive any social assistance (from 7 to 15 percent between rounds 1 and 3). The largest change was observed in Kampala where the share of households without any assistance almost tripled from (31 percent in round 1 to 86 percent in round 3). ✓ Despite lower incidence of social assistance, for those households who did receive it, more often indicated that their levels of income were equal or close to their income levels before March 2020. Access to basic needs: ✓ About 36 percent of refugee households did not have access to medicine when needed and were unable to access sufficient amounts of drinking water. In both cases, Somali refugees in the South West region had the lowest access rates. ✓ The overall share of refugee households who were not able to access medical treatment when needed dropped from 25 percent in round 2, to 18 percent in round 3. Access improved in the South West and West Nile regions and declined in Kampala. ✓ Kampala was the only region where the inability to buy main staple foods increased in round 3 compared to round 2. In round 3, 55 percent of refugees in Kampala reported being unable to buy main staple foods, compared to 26 percent in the South West and 28 percent in the West Nile regions. Knowledge and behavior: ✓ The reduction in selected safe practices slowed down by February/March 2021 relative to round 2. Refugees in Kampala more often reported to have used safe practices when compared to refugees in other regions. 7 ✓ Self-reported high levels of mask wearing while in public among refugee respondents (93 percent) may reflect over-reporting. When asked about other people using masks, refugees reported that on average only 7 out of 10 people they knew wore masks all or most of the time in public. This may be indicative of lower real compliance. ✓ Acceptance to be vaccinated and willingness to be COVID-19 tested for free were very high among refugees in the third round. At the same time, the real incidence of laboratory testing for COVID- 19 remained low (3 percent at the national level) with the highest prevalence in Kampala (8 percent). Socio-economic shocks and coping strategies: ✓ The intensity of shocks declined gradually across rounds from 38 percent of refugee households experienced multiple shocks in round 1 compared to only 17 percent in round 3. Nevertheless, every refugee household experienced at least one shock in the month preceding the interview conducted in February/March 2021. ✓ An increase in the prices of major food items remained the most frequent shock in round 3. ✓ Reducing food consumption remained the most common coping strategy and did not decline despite the lower number of shocks in round 3. Such a negative outcome is related to the fact that the share of households that used this strategy more than doubled in Kampala and remained stagnant in other regions. Concerns, mental health, and interactions with hosts: ✓ Patient Health Questionnaire-8 (PHQ-8) was administered to screen and measure the severity of depression. About 54 of refugee respondents reported depression, with refugee women and those in the West Nile having the highest incidence. ✓ Depression had a substantial negative impact on respondents’ lives. For more than half of refugee respondents who had at least one depression symptom, it made life (work, taking care of things at home and getting along with other people) extremely or very hard. ✓ Fewer refugee respondents perceived the coronavirus pandemic as a substantial threat for their household’s finances in round 3 compared to round 2 (59 versus 69 percent, respectively). However, concerns remained very high in Kampala (87 percent), South West (77 percent) and lower in West Nile (43 percent). ✓ About 13 percent of refugee respondents reported having had COVID-19 (confirmed or suspected) in rounds 2 and 3. About 14 percent of refugee respondents reported to know someone who had been infected during the same period. ✓ Fewer refugees felt unsafe at home in round 3 compared to round 2 (16 versus 25 percent, respectively). Refugees from Burundi were more likely to feel unsafe compared to refugees from other countries of origin. 8 Box 1: Selected findings from comparing refugees and Ugandans ✓ While employment rates and family businesses did not recover among refugees, improvements were observed among Ugandans. ✓ Fewer refugee households than Ugandans households reported key income levels to be higher or equal to the pre-COVID-19 levels. ✓ Exposure to shocks was much higher among refugees. Every refugee household experienced at least one shock between March and October/November 2020 compared to 42 percent of households among Ugandans who did not experience any shock between March and June 2020. ✓ Using savings as a coping strategy was frequently used by Ugandans (42 percent) and was not accessible to refugees (6 percent in round 1). ✓ Almost twice more refugee households had to borrow money to face the COVID-19 emergency compared to Ugandans. ✓ Food insecurity among refugees was much higher than among Ugandans (64 versus 9 percent, respectively). ✓ Refugees had lower access to sufficient drinking water and medicine when needed compared to Ugandans. ✓ There was a decline in the perceived threat of the COVID-19 outbreak for households’ finances among both refugees and Ugandans. However, more refugees than Ugandans still perceived it as a threat. ✓ There was a gradual decline in the share of respondents who avoided groups of more than 10 people with higher rates of safe behavior adopted by refugees than by Ugandans. ✓ More than half of refugee respondents reported depression compared to only five percent of Ugandans. 9 Table 1. Selected socio-economic indicators among refugees and Ugandans across different rounds Refugees Ugandans Pre- First round: Third Pre- First Fifth round: March Oct/Nov round: March round: February 2020 2020 Feb/March 2020 June 2020 2021 2021 Labor market (respondents) Employment rate, % of respondents 56 36 32 86 70 86 Open family business, % of households 37 23 48 26 41 Household income compared to pre-COVID levels (above or equal) Farm income, % 44 61 Family business income, % 32 53 Wage income, % 43 62 Exposure to shocks (households)* No shock, % 0 42 One shock, % 62 36 More than one shock, % 38 23 Coping with shocks (households experienced shocks) * Relied on savings, % 6 43 Had to borrow to face the COVID-19 emergency (households)** Had to borrow 50 23 Food security (households) Households ran out of food, % 85 65 24 9 Lack of access to selected basic needs Sufficient drinking water, % 36 2 Medicine, % 36 18 Concerns (respondents) Substantial threat from COVID-19 to households’ finances, % 73 59 67 41 Safe behavior (respondents) Avoided groups of more than 10 people, % 83 56 91 41 Mental health (respondents) Respondents with depression, % 54 5 Source: URHFPS rounds 1 and 3 and UHFPS rounds 1 and 5, authors’ calculations. Note: *For these indicators, recall period was different across URHFPS (March-October/November 2020) and UHFPS (March-June 2020). ** Recall period was different across URHFPS (March-October/November 2020) and UHFPS (March- Juley/August 2020). 10 A. Background 1. The High-Frequency Phone Survey for refugees in Uganda (URHFPS) tracked the socioeconomic impacts of the COVID-19 crisis on refugees throughout three rounds. The World Bank (WB) in collaboration with the Uganda Bureau of Statistics (UBOS) and the United Nations High Commissioner for Refugees (UNHCR) launched and conducted the URHFPS. The URHFPS tracked the impacts of the pandemic between October 2020 and March 2021. Data collection for the first round of the URHFPS took place between October 22 – November 25, 2020, the second round took place between December 5-24, 2020, and the final and third round was conducted between February 8-March 14, 2021. This brief discusses the results from the third round. Where possible and appropriate, the results are compared across the three rounds and also benchmarked against Ugandans by using the national High-Frequency Phone Survey on COVID-19 (UHFPS). Detailed results for the first round are available in Atamanov et al. (2021a) and for the second round in Atamanov et al. (2021b).1 2. The survey sample includes respondents with active phone numbers that were selected randomly from the Profile Global Registration System (ProGres) of UNHCR, and the refugee household survey carried out by UBOS and the World Bank in 2018.2 The targeted sample for the first round included 2,100 households and was representative at seven strata constructed as a combination of regions and different countries of origin: Kampala-Somalia, Kampala-other (Burundi, DRC, South Sudan), South West- Burundi (SW-Burundi), South West-DRC (SW-DRC), South West-South Sudan (SW-South Sudan), South West-Somalia (SW-Somalia), and West Nile-South Sudan (WN-South Sudan).3 The realized sample after the first round was 2,010 households. The second round of the phone survey tried to re-contact households from the first round and managed to reach 1,741 of them. In order to reduce the impact of attrition on the survey results, additional households were sampled, bringing the overall realized sample to 1,852 observations for the second round.4 In the third round, it was possible to re-contact 1,636 households from the second round and 349 households were contacted for the first time making the total third sample equal to 1,985 households. In order to reduce the bias related to only interviewing households with phone numbers and non-response, the data from the 2018 representative refugee household survey was used to produce and calibrate the weights for all three rounds of the phone survey. 1 Results from the first and second rounds in this brief may differ from those reported in Atamanov et al. (2021ab) due to corrections to demographic information. 2 World Bank (2019). 3 Selected country of origins accounted for about 97 percent of all refugees in Uganda in 2020. 4 Upcoming holidays constrained the team from increasing the realized sample further. 11 B. Economic activities a. Labor market and farm activities 3. Employment among refugee respondents continued to fall in round 3. Such a drop was driven by seasonal job stoppages in the West Nile region. Employment rates among refugee respondents declined from 43 percent in October/November 2020 to 32 percent in February/March 2021, widening the gap with the pre-lockdown rate by 24 percentage points (Figure 1). This contrasts to national employment rates, which fluctuated between 85 and 89 percent during the same period—returning to pre-COVID-19 levels. The observed decline in refugees’ employment rates in February/March 2021 was driven by the West Nile region, where the absolute majority of work stoppages were related to seasonal changes. This was in sharp contrast to work stoppages in round one (October/November) when the main reasons for work stoppages were related to business and government being closed due to COVID-19 restrictions, mainly in Kampala and the South West region (Figure 2). In contrast to the West Nile, employment among refugees in Kampala and the South West region in the third round improved vis-à-vis the second one. Figure 1. Employment status of refugees (in last 7 Figure 2. Key reasons for work stoppages among the days) across regions before the lockdown and in refugee respondents across regions in rounds 1 and 3, rounds 1-3, (% of all respondents) (% of the respondents who stopped working) Source: URHFPS 1-3 rounds, authors’ calculation. Note: Figure 2 covers the same 1608 respondents participating in all 3 rounds. Figure 1 was also constructed using a full panel – results are qualitatively the same. 4. The share of refugee households who reported owning family businesses dropped again in the third round, albeit to a lesser degree than the previous rounds. Ownership of family businesses among refugee households dropped from 27 percent in December 2020 to 23 percent in February/March 2021 (Figure 3) – driving an even larger gap relative to the pre-lockdown rate of 37 percent. As with the employment rate, the decline was again driven by the West Nile region. In contrast, although ownership of family business among Ugandans did not return to pre-lockdown levels, it steadily increased across rounds. 5. The continued decline in family business ownership among refugee households was also accompanied by a slightly growing share of households with declining or complete loss of revenues. Revenues from family businesses fell or were completely lost for 56 percent of refugee households in February/March 2021 (Figure 4). This was slightly higher by five percentage points than the reported changes in family business revenues in December 2020, when revenues were compared to the pre-COVID- 19 lockdown level. 12 Figure 3. Ownership of family business among refugee Figure 4. Changes in business revenues for rounds 2 households in rounds 1-3, (% of households) and 3 (% of households with family business) Source: URHFPS 1-3 rounds, authors’ calculation. Source: URHFPS 1 and 3 rounds, authors’ calculation. b. Changes in household income since the lockdown 6. For most refugee households income levels remained below the pre-March 2020 levels in the third round, with some positive changes observed in the West Nile and South West regions. Refugee households were asked to compare current income levels to the pre-March 2020 level. Compared to round 2 (December 2020), in round 3 (February/March 2021) more households reported that income from key income sources increased or remained the same as before March 2020 (Figure 5). A sharp increase was noted for humanitarian assistance, with 26 percent of households receiving higher or the same level of assistance as before March 2020 in round 3, compared to only 5 percent reported in round 2. However, most households still had a lower income level than before March 2020. Income levels have improved for key income sources of refugee households residing in the West Nile (despite a seasonal drop in employment rates among respondents) and the South West regions (Figure 6). Income levels did not change for refugees in Kampala, except for a slight increase in income from humanitarian assistance. In the West Nile region, income levels remained the closest to the pre-COVID-19 level. Figure 5. Households with higher or the same level of Figure 6. Households with higher or the same level of key income sources in rounds 2 and 3 compared to key income as before the outbreak in rounds 2 and 3 the averages before the outbreak (% of refugee across regions (% of refugee households) households) Source: URHFPS 2 and 3 rounds, authors’ calculation. Note: Regional results in figure 6 should be treated with caution given small sample size, particularly in West Nile. 13 7. Lower shares of refugees than Ugandans had income levels recovered to the pre-COVID levels. Despite improvement in income levels reported by refugees, higher shares of Ugandans reported income levels being the same or higher than the pre-COVID ones. Thus, 44 percent of refugee households reported farm income to be above or at the same level as the pre-March 2020 level compared to 61 percent Ugandan households. About 32 percent of refugee households reported farm income to be above or the same as the pre-March 2020 level compared to 53 percent Ugandan households. Finally, 43 percent of refugee households reported wage income to be above or the same as the pre-March 2020 level compared to 62 percent Ugandan households. c. Impact on poverty and food security 8. As of February/March 2021, the estimated consumption poverty among refugees has not returned to the pre-COVID level. The consumption poverty model was built by Figure 7. Predicted poverty rates before COVID-19 and using the representative household survey of across rounds using SWIFT methodology (% of refugee refugees and host communities conducted in population) 2018 to identify the strongest correlates of consumption. Consumption was imputed to refugee households participating in the URHFPS in each round. Poverty increased by seven percentage points from 44 percent before March 2020 to 51 percent in October/November 2020 and declined only slightly in the next four months, still more than four percentage points from the pre-COVID level (Figure 7). Source: URHFPS 1-3 rounds, authors’ calculation. 9. Even though the frequency of the most severe forms of food insecurity declined slightly in the third round, refugees continue to face very high levels of food insecurity. There were fewer households in Figure 8. Food security during last 30 days in rounds 1-3, (% of February/March 2021 in which adults refugee households) experienced the most severe forms of food insecurity during the last 30 days, such as going without eating for a whole day and not eating even when hungry, compared to December 2020 (Figure 8). There were no noticeable differences about the uncertainty of food supply and sufficient variety of foods. For one of the food security indicators, “household ran out of food�, the situation continued improving but still remained worse than what was reported in a comparable question in the 2018 refugee survey. This indicator was also almost seven times 14 higher than what Ugandans reported in Source: URHFPS 1-3 rounds, 2018 refugee survey, authors’ calculation the round 5 of the UHFPS (February and UHFPS round 5. 2021) – about 9 percent of households ran out of food. 10. Improvements in food security observed at the national level only held in the South West and to a lesser extent in the West Nile region. Improvements with regards to the most severe forms of food insecurity were most pronounced in the South West region (Figure 9). There was improvement in the West Nile region as well, but to a lesser extent. However, in Kampala the situation did not improve and refugees there, were the most food insecure. Figure 9. Food security during last 30 days across regions in rounds 1-3, (% of refugee households) a) Kampala a) South West a) West Nile Source: URHFPS 1-3 rounds, authors’ calculation. d. Social assistance 11. The incidence of social assistance declined across rounds, mainly in Kampala. There was a slight decline in the incidence of social assistance across rounds at the national level, masking variations at the regional level (Figure 10, Figure 11). Cash transfers remained the key form of social assistance in the South West region, with a decline in the incidence of food and other in-kind assistance. Food aid remained widespread in the West Nile region with a slight decline in other in-kind forms of aid. The largest difference was observed in Kampala, where the share of households without any assistance almost tripled between October/November 2020 and February/March 2021. Figure 10. Types of social assistance in rounds 1-3 (% of Figure 11. Types of social assistance across refugee households) regions in rounds 1-3, (% of refugee households) Source: URHFPS 1-3 rounds, authors’ calculation. 15 C. Knowledge and behavior 12. The reduction in selected safe practices to prevent the transmission of COVID-19 slowed down by February/March 2021, with refugees in Kampala still reporting that they followed safe practices most diligently. Compared to round 2 (December 2020), in round 3 (February/March 2021), refugees did not report a significant reduction in following safe practices such as avoiding handshakes, physical greetings and large groups of people (Figure 12). Refugees in Kampala were more likely to follow safe practices (Figure 13). Refugees in the South West were least likely to avoid groups of more than 10 people, while in the West Nile region refugees were the least likely to avoid handshakes and physical greetings. These regional differences are consistent with the frequency of attending religious and social gatherings in round 3. Further, more than half of refugees in Kampala did not attend any gatherings compared to 19 percent in the South West and 23 percent in the West Nile regions (Figure 15). Figure 12: Prevalence of safe practices in rounds 1 Figure 13: Prevalence of safe practices across regions (since March), rounds 2 and 3 (last week), (% of in rounds 1 (since March), rounds 2 and 3 (last week), respondents) (% of respondents) Source: URHFPS 1-3 rounds, authors’ calculation. Figure 14: Frequency of mask wearing when in public, in Figure 15: Attendance of religious or social gatherings rounds 2 and 3 (%) and number of known people who during last 7 days in round 3, (% of respondents) wear masks most or all of the time when in public, in round 3 Source: URHFPS 2 and 3 rounds, authors’ calculation. Source: URHFPS 3 round, authors’ calculation. 16 13. Refugees continue to report widespread wearing of masks while in public, but there is some evidence that this may be substantially overreported. When asked about wearing masks while in public, more than 90 percent of refugees reported doing this all or most of the time (Figure 14). However, when respondents were asked how many people out of ten, they knew who wore masks all or most of the time in public, the average number reported was only seven people. Roughly speaking, this is equivalent to 70 percent of people wearing masks in public. The discrepancy between the two indicators, coupled with anecdotal feedback from UNHCR may signal that self-reported mask wearing is over-reported and real compliance may be significantly lower. 14. Acceptance to be vaccinated and willingness to be COVID-19 tested for free were very high among refugees. Acceptance to be vaccinated if a free and safe vaccine were available remained at around 90 percent in the third round (Figure 16). This is comparable to the acceptance rate among Ugandans in February 2021 equal to 88 percent. Refugees of Burundi origin have the lowest, but still high, acceptance rate (80 percent). Respondents were also asked if they would be willing to be COVID-19 tested for free and more than 90 percent of them agreed (Figure 17). This is close to what was reported by Ugandans in February 2021 (93 percent). At the same time, the real incidence of laboratory testing for COVID-19, which was not free, remained quite low among refugees in round 3. It was about 3 percent at the national level with the highest prevalence in Kampala (8 percent). Figure 16: Acceptance to be vaccinated if approved Figure 17: Household with any member ever received a vaccine was available right now at no cost in rounds 2 laboratory diagnosis for COVID-19 and respondents who and 3, (% of refugee respondents) would be willing to be tested at no cost in round 3, (%) Source: URHFPS 3 round, authors’ calculation. 17 D. Access to goods and basic needs 15. Refugee households had the lowest access to medicine and sufficient drinking water in round 3. Respondents were asked about access to medicine when needed during the seven days preceding the interview. Overall, 98 percent of households needed medicine and among them 36 percent were not able to access it. This share was the highest among Somali refugees living in Kampala (54 percent)5 and Somali refugees living in the South West region (57 percent; Figure 18). Respondents were also asked if there was any time during the week preceding the interview when they did not have sufficient drinking water. About 36 percent of households experienced this in the third round with the highest share observed among Somali refugees in the South West region (66 percent). 16. Nationally, access to medical treatment improved by the third round, but there was a significant decline in Kampala. The share of refugee households who were not able to access medical treatment when needed dropped from 25 percent in round 2 to 18 percent in round 3 (Figure 19). This positive trend is almost fully attributable to the West Nile region. However, in Kampala, there was a substantial increase in the share of households who could not access medical treatment, from 28 percent in round 2 to 43 percent in round 3. Figure 18: No access to medicine during last month Figure 19: No access to medical treatment when needed when needed by regions in round 3 (% of refugee by regions and across rounds 1-3 (% of refugee households) households) Source: URHFPS 3 round, authors’ calculation. Source: URHFPS 1-3 rounds, authors’ calculation. Note: Recall period in round 1 was since March 2020, for round 2 and 3 last month. 5 Refugees in Kampala do not have access to free medicines the same way as refugees in settlements. 18 17. Overall, the frequency of households Figure 20: Inability to buy main staple food when needed by being unable to buy main staple foods also regions in rounds 1-3, (% of refugee households) improved between rounds 2 and 3, with Kampala being an exception. The share of refugee households who were not able to buy main staple foods when needed dropped between rounds 2 and 3 (38 percent in round 2 to 28 percent in round 3). However, households in Kampala experienced more difficulties in buying food in the third round which was probably a reflection of income constraints/high food prices rather than food availability (Figure 20). Source: URHFPS 1-3 rounds, authors’ calculation. 18. Compared to Ugandans, refugees’ ability to buy food, access sufficient drinking water, and access medicine and medical treatment were significantly lower. Figure 21 compares access to different basic goods and services among Ugandans and refugee households using similar time periods to the extent possible. Even though definitions are not fully comparable for all indicators, the difference between refugees and Ugandans is stark for selected basic goods and services. About 36 percent of refugee households did not have sufficient drinking water to meet household needs at least once in the last seven days. Only two percent of Ugandans face scarcity in drinking water. About 36 percent of refugee households were not able to buy medicine when needed compared to 18 percent among Ugandans. The same pattern is observed for inability to buy food and access to medical treatment. Only access to masks was better among refugees than among Ugandans. However, only a very small share of households reported being unable to access them. Figure 21: Share of households unable to cover basic needs and access services among refugees and Ugandans (% of households) Source: URHFPS 1-3 rounds, UHFPS 3-5 rounds, authors’ calculation. Note: Definitions of some indicators differ slightly across the refugee and national phone surveys. In addition, survey periods are also different for some indicators. Thus, ability to buy food among Ugandans is measured in September/October 2020, ability to access soap, water to wash hands, sufficient drinking water and medical treatment in October/November 2020, and access to masks, medicine, and medical treatment in February 2021. 19 E. Coping strategies to socioeconomic shocks 19. In all rounds, every refugee household experienced at least one shock, with the number of refugees facing multiple shocks declining overall and particularly in the South West and West Nile regions. All refugee households experienced at least one shock during the month preceding the interview in the third round (Figure 22). However, the share of households who experienced multiple shocks declined significantly from 38 percent in round 1 to 17 percent in round 3. This positive tendency was observed in the South West and West Nile regions, while in Kampala, the share of households who experienced multiple shocks increased in round 3 compared to round 2 (Figure 23). Among Ugandans though, about 42 percent of households did not experience any shock during March-June 2020 period. Figure 22: Number of shocks in rounds 1-3, (% of Figure 23: Households experienced multiple shocks, (% refugee households) of refugee households) Source: URHFPS 1-3 rounds, authors’ calculation. 20. As for rounds 1 and 2, the most Figure 24: Type of shocks in, (% of refugee households) frequent shock in round 3 was an increase in the price of major food items. The incidence of different types of shocks was less prevalent in round 3, compared to rounds 1 and 2. The only exception was illness, injury or death of income earning member of household (Figure 24). This type of shock became the second most common shock in round 3, with shocks associated with an increase in the prices of major food items being the most common. Source: URHFPS 3 round, authors’ calculation. 20 21. Reducing food consumption Figure 25: Selected strategies to cope with socioeconomic shocks remained the most frequently used coping in rounds 1-3, (% of refugee households) strategy and did not decline despite the lower average number of shocks in round 3. About 24 percent of refugee households who experienced shocks in round 3 used the negative coping strategy of reducing food consumption (Figure 25). This share did not fall despite the lower average number of shocks in round 3. Such a negative outcome is related to the fact that the share of households that used this strategy more than doubled in Kampala, while remaining stagnant in other regions. Compared to Ugandans, refugees were less likely to use savings as a coping strategy: 6 percent of Source: URHFPS second round, authors’ calculation. refugees in round 1 vis-à-vis more than 43 percent among Ugandans between March- June 2020. F. Mental health, concerns and interactions with hosts 22. More than half of respondents reported suffering from depression in the third round, with a higher incidence among women refugees, elderly, and those living in the West Nile region. The Patient Health Questionnaire-8 (PHQ-8) was administered to screen and measure severity of depression among respondents. PHQ-8 was used in the national and refuge HFPS. The tool rates the frequency of depression symptoms, adds them up to a severity score and measures the prevalence of depression. About 54 percent of refugee respondents reported depression in the third round which is more than 10 times higher than the reported rate of depression among the Ugandans (Figure 26). Incidence of depression was extremely high among refugee respondents in the West Nile (71 percent). Women respondents and those above 60 years old also reported higher depression rates. Refugees are generally more likely to face depression and post-conflict post-traumatic stress disorder (Karam et al. 2014, Steel et al. 2009), which has likely been exacerbated by the pandemic. Figure 26: Respondents with major depression, (% of respondents) Source: URHFPS 3 round and UHFPS 5 round, authors’ calculation. Note: Respondents with the score 10 and above are considered to have a major depression. 21 23. High depression rates among refugees had significant negative impacts on individuals’ capacity to work, do housework and engage in social interactions. Those refugees, who indicated that Figure 27: Impact of depression on work, housework or getting along they had experienced at least one with other people, (% of refugee respondents who experience at least depression symptom for more than one depression symptom in two weeks) one day during the last two weeks were asked about the impact of such symptoms on their lives (work, housework or social interactions). For more than half of these refugee respondents, depression symptoms have made components of their lives extremely or very hard (Figure 27Figure 27). Refugee women and those without formal education were more likely to report that components of their lives have become very or extremely hard due to depression symptoms. Source: URHFPS 3 round, authors’ calculation. 24. Fewer refugee respondents perceived the coronavirus pandemic as a substantial threat to households’ finances in West Nile in round 3, but concerns remained quite high in Kampala and the South West. In about 59 percent of refugee households, Figure 28: Perceived substantial threat from coronavirus respondents considered the pandemic as a pandemic to household’s finances, (% of refugee households) substantial threat to their households’ finance in round 3. This was lower than in rounds 2 and 1, but still substantially higher than what Ugandans reported in February 2021 (41 percent; Figure 28). The decline was driven by the West Nile region, while in Kampala and South West the share of perceived concern either increased or did not change. Source: URHFPS 1-3 rounds, authors’ calculation. 25. About 13 percent of refugee respondents reported to have had COVID-19 and knew someone who was infected in rounds 2 and 3. Combining information from rounds 1 and 2, 13 percent of respondents reported to have had COVID-19 (suspected or confirmed, Figure 29). This was higher than seven percent reported by Ugandans in February 2021. Suspected incidence of COVID-19 was much higher in the West Nile region and among men. About 14 percent of respondents knew somebody from their immediate social environment who has had COVID-19 (suspected or confirmed). This was lower than 19 percent reported by Ugandans in February 2021. The largest share was observed in the South West (15 22 percent) even though only five percent of respondents in this region suspected to have had COVID-19 themselves. Figure 29: Respondents who believed themselves to Figure 30: Respondents who did not feel safe at home, have had COVID-19 and those who knew someone who (% of respondents) had it, cumulative from rounds 2 and 3, (% of respondents) Source: URHFPS rounds 2 and 3, authors’ calculation. Source: URHFPS round 3, authors’ calculation. 26. Refugee respondents felt safer at home in round 3 compared to the previous rounds, with refugees from Burundi feeling unsafe the most. Only 16 percent of respondents felt unsafe at home in the third round (Figure 30). This was much lower compared to rounds 1 and 2 (22 and 25 percent accordingly). Refugees from Burundi were more likely to report feeling unsafe at home (41 percent) compared to only seven percent of Somali respondents. More men respondents felt unsafe compared to women (21 versus 12 percent accordingly). 23 G. Bibliography Atamanov, A., Yoshida, N., Beltramo, T.P., Rios Rivera, L.A., Sarr, I., Waita, P. and Yoshimura, K. (2021a). Monitoring Social and Economic Impacts of COVID-19 on Refugees in Uganda: Results from the High- Frequency Phone Survey - First Round (English). Monitoring COVID-19 Impacts on Refugees in Uganda Washington, D.C.: World Bank Group. Available from: http://documents.worldbank.org/curated/en/682171613766616044/Monitoring-Social-and-Economic- Impacts-of-COVID-19-on-Refugees-in-Uganda-Results-from-the-High-Frequency-Phone-Survey-First- Round. Atamanov, A., Yoshida, N., Beltramo, T.P., Rios Rivera, L.A., Sarr, I., Waita, P. and Yoshimura, K. (2021b). Monitoring Social and Economic Impacts of COVID-19 on Refugees in Uganda: Results from the High- Frequency Phone Survey – Second Round (English). Monitoring COVID-19 Impacts on Refugees in Uganda Washington, D.C.: World Bank Group. Available from: http://documents.worldbank.org/curated/en/377241617082682333/Monitoring-Social-and-Economic- Impacts-of-COVID-19-on-Refugees-in-Uganda-Results-from-the-High-Frequency-Phone-Survey-Second- Round. Karam, Elie G., Matthew J. Friedman, Eric D. Hill, Ronald C. Kessler, Katie A. McLaughlin, Maria Petukhova, Laura Sampson, et al. “Cumulative Traumas and Risk Thresholds: 12-Month PTSD in the World Mental Health (WMH) Surveys.� Depression and Anxiety 31, no. 2 (February 2014): 130–42. https://doi.org/10.1002/da.22169. Steel, Zachary, Tien Chey, Derrick Silove, Claire Marnane, Richard A. Bryant, and Mark van Ommeren. “Association of Torture and Other Potentially Traumatic Events With Mental Health Outcomes Among Populations Exposed to Mass Conflict and Displacement: A Systematic Review and Meta-Analysis.� JAMA 302, no. 5 (August 5, 2009): 537. https://doi.org/10.1001/jama.2009.1132. Yoshida, N., Munoz, R., Skinner, A., Kyung�eun Lee, C., Brataj, M., Durbin, W. and Sharma, D. (2015). SWIFT Data Collection Guidelines. Version 2. Washington, DC: World Bank. World Bank (2019). Informing the Refugee Policy Response in Uganda: Results from the Uganda Refugee and Host Communities 2018 Household Survey (English). Washington, D.C.: World Bank Group. Available from: http://hdl.handle.net/10986/32511. World Bank. (2020). COVID-19 Impact Monitoring: Uganda, Round 3. World Bank, Washington, DC. 24 H. Annex a. Sampling The Profile Global Registration System (ProGres) served as a sample frame for the URHFPS. It was complemented by the data collected for the refugee household survey carried out by UBOS and the World Bank in 2018. The sample was selected from the pool of refugees with phone numbers. The targeted sample included 2,100 observations: 300 observations in each stratum. Four countries of origin were targeted in the survey: Burundi, Democratic Republic of Congo (DRC), Somalia and South Sudan. The combination of country of origin and region were used to create seven strata: Kampala-Somalia, Kampala- other (Burundi, DRC, South Sudan), South West-Burundi, South West-DRC, South West-South Sudan, South West-Somalia, and West Nile-South Sudan. The realized sample of data collection was 2,010 households in the first round and 1,852 households in the second round. The number of observations in the first two rounds and population shares across strata are shown in Table 1. Given that the refugee population with phone numbers may be fundamentally different from the population without phone numbers (e.g. more educated, affluent, living in urban areas and so forth), and also due to the issue of non-response, respondents in the phone survey may not be representative of the overall refugee population in Uganda. In order to reduce the potential bias and make results nationally representative, we developed a weighting procedure by producing weights using a nationally representative survey as a reference and calibrating obtained weights to make the phone survey nationally representative and resemble the distribution for the key variables of those from the reference survey. The refugee survey from 2018 was used as a reference in creating weights for the phone survey, while post-stratification was conducted to preserve to the extent possible regional population shares from the ProGres dataset as of November 2020. Table 1: Distribution of observations and population across strata in rounds 1, 2 and 3 Round 1 Round 2 Round 3 number number Population Population number of Population of of Strata weighted weighted observatio weighted observa observati share share ns share tions ons Kampala (Burundi, DRC, South Sudan) 292 3 269 3 286 3 Kampala-Somalia 340 2 300 2 320 2 SW-Burundi 284 3 271 3 300 3 SW-DRC 299 28 305 28 315 27 SW-South Sudan 312 5 293 5 299 5 West Nile-South Sudan 245 58 201 58 233 58 SW-Somalia 238 1 213 1 232 1 Total 2,010 100 1,852 100 1,985 100 Source: URHFPS 1-3 rounds, authors’ calculation. 25 b. Profile of respondents and households Table 2. Characteristics of respondents and households in the second round Residence Overall Kampala South West West Nile Demographics Household size 5.4 5.2 5.0 5.8 Number of children, 0-14 3.0 2.2 2.8 3.3 Dependency ratio, % 54 37 53 57 Age of respondent 38 37 39 37 Male respondent, % 48 42 60 40 Male head, % 49 45 65 38 Education of respondent No formal education, % 21 27 26 16 Complete primary and below, % 37 16 36 39 Some secondary, % 31 22 23 39 Completed secondary & +, % 12 35 15 7 Year of arrival Before 2011, % 13 18 28 2 2011-2017, % 78 61 61 92 2018-2020, % 9 22 11 6 Country of origin South Sudan, % 61 5 16 100 DRC, % 27 40 61 0 Burundi, % 6 8 13 0 Somalia, % 6 48 10 0 Source: URHFPS 3 round, authors’ calculation. Note: All estimates are household weighted. In third round information about education of respondent was not collected and reported numbers are only for respondents participating in rounds 2 and 3. 26