REPORT NO. 4 MARCH 2021 Results from the High-Frequency Phone Survey of Households Round 4 (17 December 2020–12 January 2021)1 Authors: Wendy Karamba, Isabelle Salcher and Kimsun Tong INTRODUCTION The coronavirus disease 2019 (COVID-19) pandemic and its effects on households create an urgent need for timely data and evidence to help monitor and mitigate the social and economic impacts of the crisis on the Cambodian people, especially the poor and most vulnerable. To monitor the evolving socioeconomic impacts of the COVID-19 pandemic and inform policy responses and interventions, the World Bank designed and conducted a nationally representative High-Frequency Phone Survey (HFPS) of households in Cambodia. The survey covers important and relevant topics, including knowledge of COVID-19 and adoption of preventative behaviors, economic activity and income sources, access to basic goods and services, exposure to shocks and coping mechanisms, and access to social assistance. The survey will track the same households over 10 months, with selected respondents—typically the household head—completing interviews every 8 weeks. Monitoring the well-being of households over time will improve understanding of the effects of, and household responses to the COVID-19 pandemic in near-real time. This brief summarizes the findings of Round 4 (R4) of the HFPS. The information presented is based on a sample of 1,687 households, of which 410 were drawn from the nationally representative Living Standards Measurement Study Plus (LSMS+) survey and 1,277 from the list of beneficiaries of the conditional cash transfer program for pregnant women and children under 2 with an IDPoor equity card (IDPoor sample). IDPoor is Cambodia’s national poverty identification program and official targeting mechanism for programs that support the poor. Unless otherwise noted, the results presented are primarily drawn from the LSMS+ sample. To trace the evolution of key indicators, the results of Round 4 (R4) implemented in December 2020–January 2021 are compared against results of Round 3 (R3) implemented October–November 2020, Round 2 (R2) implemented August–September 2020, and Round 1 (R1) implemented May–June 2020. HIGHLIGHTS -- HFPS ROUND 4 Access to food staples remained robust as markets continued to function well. Almost all households seeking to buy medicines or access ❖ medical treatment were able to do so, even as demand for health services remained high. ❖ Children’s engagement in learning activities declined due to the school term holiday and the second nationwide closure of schools on November 30, 2020 following a local COVID-19 outbreak. The proportion of households with school-age children ages 6–17 engaged in learning activities declined from 92 percent in October—where it had reached pre-pandemic levels—to 57 percent in December 2020. Unlike the first nationwide school closure where instruction shifted to remote settings while the academic term progressed, the second school closure cancelled the remainder of the 2019–2020 academic year for public schools and suspended in-person instruction for private schools for 2 weeks while e-learning was applied during this period. Children returned to using remote alternatives such as mobile learning applications and fewer had face-to-face meetings with teachers. There was greater disengagement in learning activities for children in IDPoor households, who are more likely to rely on public education. 1 Kimsun Tong and Wendy Karamba led the Cambodia High-Frequency Phone Survey (HFPS) team that comprised of Maheshwor Shrestha, Sokbunthoeun So, and Isabelle Salcher. Nuppun Research Consulting implemented the survey with technical and financial support from the World Bank. Additional contributions for the HFPS were received from the Public Financial Management and Service Delivery Trust Fund contributed by Australia and the European Union. The Food and Agriculture Organization (FAO) of the United Nations assisted with the analysis of the food insecurity based on the application of the Food Insecurity Experience Scale (FIES). The team benefitted from advice and comments from World Bank Group colleagues including Rinku Murgai, Claire Honore Hollweg, Maheshwor Shrestha, and Simeth Beng. The team is also grateful for comments from the Ministry of Economy and Finance. REPORT NO. 4 MARCH 2021 72 percent of respondents were employed in December, thereby reaching similar levels as in May and August 2020 but remaining below its ❖ pre-pandemic level when 82 percent of respondents were working. Seasonality is the primary reason earners stopped working. The negative impacts of the COVID-19 pandemic on non-farm family businesses remain substantial with weak consumer demand driving ❖ the losses in business revenues. In December, more than half of households operating non-farm businesses (58 percent) reported having made “less” or “no revenues” relative to the previous month. Although there was a significant reduction in the share of non-farm business households reporting having made “less” or “no revenues” since May (81 percent), there was no further reduction since October (54 percent). While the share of households reporting declines in household income have slowed since May 2020, about 1 in 2 households continued to ❖ report that their household income had declined relative to the last survey. Slowing declines in labor income and the provision of the government's cash relief transfers to assist poor and vulnerable households during COVID-19 appear to have helped mitigate some of the losses. Since the launch of the nationwide COVID-19 relief transfer program in June 2020, there has been a marked increase in the share of IDPoor ❖ households receiving social assistance from the government, mostly in the form of cash transfers. By December 2020, 93 percent of eligible IDPoor households had received relief cash transfers and most of them had received multiple payments. While an increasing share of households received government assistance since May 2020, an increasing share of households have had to ❖ borrow, delay payment obligations and take on additional income-generating activities to cope with the COVID-19 crisis. Around 60 percent of households perceive that their current well-being and economic status is lower compared to the previous year, but ❖ fewer report a decline relative to June 2020–when the government expanded its social assistance program. Beneficiary households of the relief cash transfers perceive the program to have positively contributed to their economic well-being. Food insecurity remained unchanged between October and December. Among the LSMS+ sample, the prevalence of moderate-or-severe ❖ food insecurity was 17 percent in December, which includes 0.5 percent of the population affected by severe food insecurity. This is compared to 17 percent of moderate-or-severe and 0.7 percent of severe food insecurity in October. Among IDPoor households, there was a statistically insignificant (at the 95 percent confidence interval) decline in the prevalence of moderate-or-severe food insecurity from 39 percent in October to 34 percent in December. ACCESS Despite the COVID-19 pandemic, markets remain functional and able to supply basic necessities. Since May 2020 TO BASIC (Round 1), almost all households regardless of poverty status and area of residence were able to buy rice, fish/meat, NECESSITIES and vegetable/fruit when needed in the 7 days preceding the surveys. Access to medicine and health services has also remained nearly universal throughout the duration of the COVID-19 pandemic. Nearly all households that tried FOOD STAPLES to purchase medicine in the 7 days preceding the surveys were able to do so, even as the share of households demanding medicines increased and remained high since May (Figure 1). Similarly, nearly all households that needed medical treatment since the previous survey rounds (or pre-COVID-19 outbreak for Round 1) were able to access medical services even as demand continued to be high (Figure 2). While the need for medical treatment is MEDICINE higher in rural areas, there are no disparities by gender of the respondents nor by poverty status of the household. Access is universally high regardless of which households required medical treatment. Page 2 REPORT NO. 4 MARCH 2021 Figure 1: Households able to buy medicine in the last 7 days 1 1 30 28 22 34 33 30 57 55 99 99 100 100 100 100 100 100 70 72 78 % 66 67 70 43 45 R1 R2 R3 R4 R1 R2 R3 R4 R1 R2 R3 R4 R1 R2 R3 R4 LSMS+ IDPoor LSMS+ IDPoor Tried to buy medicine Able to buy medicine (for those who tried to buy) Yes No Figure 2: Households needing and able to access medical treatment since the last round 1 1 2 45 39 38 44 34 35 60 74 99 100 100 99 100 98 100 100 % 55 61 62 56 66 65 39 26 R1 R2 R3 R4 R1 R2 R3 R4 R1 R2 R3 R4 R1 R2 R3 R4 LSMS+ IDPoor LSMS+ IDPoor Need medical treatment Access medical treatment (for those who needed treatment) Yes No SCHOOLING A second nationwide closure of schools was put in effect in November in response to a local COVID-19 outbreak. The second nationwide school closure came shortly after all schools had reopened for in-person instruction. Schools throughout Cambodia reopened on November 2, 2020 for the first time since March when the first nationwide school closure was instituted. Soon after, on November 30, 2020, all schools across the country were shuttered according to a new directive of the Ministry of Education, Youth and Sport (MoEYS) issued on November 29, 2020 to curb a local outbreak of COVID-19. Public schools concluded the 2019–2020 academic year early while private schools suspended in-person instruction for 2 weeks while applying e-learning during this period. Grade 12 classes—that were preparing students to sit for the national high-school examination in December—could continue depending on the level of risk for COVID-19 transmission the high school faced. Following a school break in December/January, public schools commenced the 2020–2021 academic year on January 11, 2021 and private schools started reopening on January 4, 2021. Before the second nationwide closure, Cambodia followed a three-phase approach to school reopening. On March 16, 2020, MoEYS directed the closure of all schools nationwide in response to the COVID-19 pandemic. Students were required to continue their studies remotely with the aid of technology while the country took a phased approach to reopening schools. After a five-month period of school closures, Phase 1 of school reopening commenced in August 2020. During Phase 1 of school resumption, a handful of international private schools located in the cities of Phnom Penh, Battambang, and Siem Reap were permitted to reopen provided they followed the highest safety standards set out by the Ministry of Health to prevent the spread of COVID-19. Phase 2 of school resumption was applied on September 7, 2020, allowing some public and private schools—kindergartens, primary and secondary schools—to resume operation. Schools from grades 9 to 12 could reopen across the country, while schools of all grade levels could only reopen in low-risk provinces. On November 2, 2020, all schools across the Page 3 REPORT NO. 4 MARCH 2021 country reopened as part of Phase 3. All public and private schools ranging from preschools to high schools were allowed to resume in-person instruction. The second school closure and school term holiday—which coincided with Round 4 of the HFPS—reduced children’s engagement in learning activities. The proportion of LSMS+ households with school-age children (ages 6–17) engaged in learning activities dropped to 57 percent in December 2020, after having risen from 62 percent in May, to 75 percent in August, and 92 percent in October during the phased resumption of in-person instruction (Figure 3). Pre-pandemic levels had been reached in October, when Phase 2 of school resumption was under effect. In December, slightly fewer households with school-age children participated in educational activities compared to May, when schools were also closed nationwide but the academic year was in progress with e-learning being applied. The closure of school reduced learning opportunities for children in IDPoor households even more dramatically, likely due to a greater reliance on public schooling which concluded early. In December 2020, only 33 percent of IDPoor households with school-age children reported that their children engaged in learning activities in the past 7 days, compared with 57 percent of LSMS+ households with school-age children (which capture the overall population of households) (Figure 3). Despite schools being closed in both periods, only half as many IDPoor households with school-age children engaged in learning activities in December than in May (68 percent). This may be due to the fact that the academic year had concluded among public schools rather than due to a lack of remote learning opportunities. Figure 3: Children engaged in education or learning activities in last 7 days With schools temporarily closed for in-person instruction, some households returned to remote learning alternatives. The proportion of households with school-age children using mobile learning applications rose by 25 percentage points since October to 35 percent in December 2020 (Figure 4). There also has been a slight increase in the proportion of households with school-age children watching educational television programs, from 3 percent in October to 7 percent in December. Fewer households met with teachers in December (Round 4) than during the phased resumption of in-person instruction (Round 2 and Round 3). In December, the proportion of households with school-age children that met with teachers was 31 percent compared with 86 percent in October and 66 percent in August. In December, IDPoor households with school-age children were equally likely to have met with teachers and watched educational television programs as LSMS+ households but were less likely to have used mobile learning applications and completed assignments. Page 4 REPORT NO. 4 MARCH 2021 Figure 4: Children engaged in education or learning activities in last 7 days 100 86 82 78 80 66 66 58 60 47 4544 36 40 31 2832 31 32 35 29 29 28 20 24 % 20 14 15 14 13 10 10 8 7 3 3 2 0 0 2 3 2 0 0 0 0 Met with Met with Listened to Listened to assignments Watched TV assignments Watched TV learning apps learning apps Completed Completed teacher teacher radio radio Mobile Mobile LSMS+ IDPoor Round 1 Round 2 Round 3 Round 4 School closures disproportionately affected learning activities of the most vulnerable students. While school closures are a logical solution to enforce social distancing within communities, they disrupted learning opportunities of children in households in the bottom 40 percent of the income distribution (“poorest”) more than in the top 60 percent (“richer”). After the participation gap in learning activities between the poorest and richer households had begun to close as schools reopened, it rose from a 6 percentage point gap in October to a 14 percentage point gap in December (Figure 5). While this could be attributed to the early conclusion of the school year due to nationwide school closures, this could also signal that children in richer households are more likely to attend private schools which had not concluded the school year. In December, school-age children in the poorest households were half as likely to have met with a teacher in the week preceding the survey than those in richer households. In contrast, the poorest households were more likely to have met with a teacher than richer households in August (Round 2) and October (Round 3). Although the poorest and richer households were equally likely to use mobile learning applications in December, usage of mobile learning applications was considerably higher among richer households since March when schools first closed at the onset of the global pandemic. In May, 66 percent of richer households with school-age children reported using mobile learning applications compared to only 48 percent of the bottom 40 percent with school-age children. Figure 5: Households with school-age children engaged in education or learning activities in last 7 days 92 92 92 95 89 86 76 80 75 69 70 66 68 64 62 62 63 64 57 57 59 52 50 % 33 All Urban Rural Bottom 40 Top 60 LSMS+ IDPoor Round 1 Round 2 Round 3 Round 4 Page 5 REPORT NO. 4 MARCH 2021 EMPLOYMENT Employment has yet to return to pre-pandemic levels. Prior to the COVID-19 outbreak 82 percent of respondents PATTERNS AND were working. Following the onset of the pandemic, 71 percent of respondents were working in May and this HOUSEHOLD remained relatively unchanged until October when there was a further decline in employment to 65 percent (Figure INCOME 6). Employment increased back to 72 percent in December. Of those currently working, 94 percent worked in the same job as in the previous round, and only 6 percent switched jobs. The share of respondents who stopped working since the last round has remained at around 10 percent since May (Figure 6). Seasonality in farming (33 percent) and seasonality in employment (23 percent) are the primary reasons earners had stopped working. Possibly due to the recent local outbreak, the share of respondents who stopped working due to COVID-19 related EMPLOYMENT business closures rose to 12 percent, reaching similar levels as in August but remaining below levels reported in STATUS May (40 percent). Employment for the main earners in IDPoor households has continuously increased since May. Seasonality was the primary reason the main earner in IDPoor households stopped working since the last round. Figure 6: Respondents (LSMS+) or household’s main Figure 7: Changes in sales revenues in non-farm earner (IDPoor) working in last 7 days household businesses relative to last month 7 4 5 5 5 5 8 5 10 18 20 21 19 16 10 10 11 5 4 15 11 10 13 9 31 6 5 1 4 8 40 37 44 42 46 % 78 82 65 65 64 68 65 % 73 61 52 58 48 48 40 R1 R2 R3 R4 R2 R3 R4 8 3 3 2 4 5 LSMS+ IDPoor Out of work since last round R1 R2 R3 R4 R2 R3 R4 Stopped working since last round LSMS+ IDPoor Currently working (changed job since last round) Currently working (same job since last round) No revenue Less The same Higher The negative impacts of the COVID-19 pandemic on non-farm family businesses remain substantial. More than 1 in 2 (58 percent) non-farm household businesses continued reporting earning “less” or “no revenues” relative to the month prior in December 2020 (Figure 7). This is compared to 81 percent in May, 64 percent in August, and 54 percent in October. Meanwhile, the share of households reporting their business revenues “stayed the same” doubled between May and December. Note, about one-third (30–32 percent) of LSMS+ households and one-tenth (12–13 percent) of IDPoor households operate a non-family business. The share of households operating a non- farm family business has remained relatively unchanged among LSMS+ and IDPoor households. However, there have been shifts in sector. After rising to around 80 percent in August and October, the share of businesses that were in retail returned to its May level of 67 percent in December. In turn, more businesses were in manufacturing and personal services. Household businesses in urban areas and in the top 60 percent of the income distribution were more likely to report less or no revenues. Weak consumer demand continues to drive revenue losses for non-farm family businesses. Having fewer or no customers, as reported by 81 percent of households operating a non-farm business, is the primary reason for not generating revenue. Lower demand is disproportionately affecting household businesses operating in urban areas, households in the top 60 percent of the income distribution, and households headed by a woman. INCOME While widespread reductions in household income continue irrespective of poverty status or area of residence, there are signs of slowdown with fewer households experiencing declines. In Round 4, about 1 in 2 households (48 percent) experienced declines in income between October and December (Figure 8). This compares to 51 percent in Round 3 that experienced declines between August and October, 63 percent in Round 2 that experienced declines between May and August, and 83 percent in Round 1 that experienced declines between the COVID-19 outbreak and May. With households reporting an average reduction in total household income of 41 Page 6 REPORT NO. 4 MARCH 2021 percent in Round 2, 39 percent in Round 3, and 40 percent in Round 4, the amount of income loss is assumed to be considerable. Fewer households report that labor income had decreased since the previous round, suggesting slowdowns in the share of households experiencing declines compared to the onset of the COVID-19 pandemic. Yet, one-third of households reported a reduction in income from wage employment or family farming between October and December, while 1 in 2 households reported lower income from non-farm family businesses. Incomes from remittances and family or non-family assistance also fell since the previous round, but a smaller share of households draw their income from these sources. Increased assistance from the government and non- governmental organizations (NGOs) also might have helped moderate losses in household income. Among households that had received such assistance in the last 12 months, 70 percent reported that their income from this source “stayed the same” between October and December in Round 4. About 70 percent also reported that income from this source “stayed the same” between August and October in Round 3. This came after 42 percent reported an increase between May and August in Round 2, which coincided with the launch of the COVID-19 relief cash transfer program. This represented a 10-fold increase from May (4 percent). Figure 8: Changes in total household income since Figure 9: Changes in total household income since the last round November 2019 1 8 7 9 1 12 9 9 16 11 22 29 34 42 43 32 45 46 27 % 88 % 83 40 63 57 51 48 44 46 51 26 R1 R2 R3 R4 R1 R2 R3 R4 LSMS+ IDPoor LSMS+ IDPoor Reduced Stayed the same Increased Reduced Stayed the same Increased A higher share of households report experiencing reductions in household income relative to the previous year in November 2019 than those that report experiencing increases. Around 1 in 2 households (51 percent) reported that their household income had declined relative to the previous year in November 2019 (Figure 9). These households reported an average reduction in income of 43 percent. In contrast, only one-quarter of households (22 percent) reported an increase in total household income since November 2019. These households reported incomes were 28 percent higher on average. For many households, household income fell largely due to shocks to labor income (farm and non-farm). Households experiencing a reduction in these income sources reported that these sources decreased by 40 to 45 percent on average since November 2019. Remittances were also negatively impacted, but the share of households sourcing their livelihoods from remittances is relatively small (10 percent). Most households derive their livelihoods from wage employment (69 percent), family farming (51 percent) and non-farm family businesses (35 percent). Among households deriving income from social assistance, there has been a striking increase in assistance received from the government or NGOs—likely following the launch of the COVID-19 relief transfer program and increased support from NGOs during the pandemic. Fewer IDPoor households reported that their income declined relative to November 2019 (Figure 9). Page 7 REPORT NO. 4 MARCH 2021 Figure 10: Changes in income by source since November 2019 IDPoor Family 48 30 22 farm LSMS+ 49 32 18 IDPoor 54 27 20 family Non- farm biz LSMS+ 70 20 10 IDPoor 54 35 11 non- Remitt- Wage NGO Pension property family ances empl. LSMS+ 44 43 13 IDPoor 74 11 15 LSMS+ 56 35 9 Income family / IDPoor Assist- 25 35 41 ance LSMS+ 28 47 24 IDPoor 26 16 58 from LSMS+ 38 40 22 IDPoor 74 18 LSMS+ 72 26 IDPoor 7 5 89 Assist- Govt/ ance LSMS+ 5 10 85 % Reduced Stayed the same Increased Perceived household well-being and economic status have deteriorated in recent months as well as relative to PRECEIVED WELL-BEING the previous year. Compared to the previous year in November 2019, 60 percent of households perceive their AND COPING well-being and socioeconomic status to have fallen, while 30 percent perceive it to have stayed the same (Figure 11). This indicates that the COVID-19 pandemic has hit households hard and that disruptions to employment and livelihoods have led to a perceived deterioration in well-being. However, compared to a more recent period in June 2020, fewer households reported a perceived reduction in well-being and socioeconomic status. Only 46 percent of households in Round 4, perceived their well-being and status to have decreased compared to June (Figure 11). On the one hand, this signals that perceived well-being and status continued to deteriorate for some households, either due to COVID-19 or other factors such as the flooding experienced in October. On the other hand, social assistance provided by the government starting in June 2020 might have helped slow down this reduction. Figure 11: Changes in household well-being and economic status Compared to November 2019 Compared to June 2020 9 15 3 9 8 13 30 29 48 45 53 49 % % 41 31 40 33 30 23 19 24 9 13 9 15 LSMS+ IDPoor R3 R4 R3 R4 LSMS+ IDPoor Increased by a lot Increased by a lot Increased by a little bit Increased by a little bit Stayed about the same Stayed about the same Reduced by a little bit Reduced by a little bit Reduced by a lot Reduced by a lot Page 8 REPORT NO. 4 MARCH 2021 Households have employed a variety of coping strategies during COVID-19, but are increasingly borrowing, delaying payments, and finding supplemental work in addition to getting social assistance. In the early period of the pandemic (Round 1), the most common coping strategies consisted in reducing food and non-food consumption, purchasing on credit, and borrowing from friends and family (Figure 12). These coping strategies were more commonly adopted among IDPoor households. For instance, 86 percent and 72 percent of IDPoor households with a valid equity card had to reduce their food and non-food consumption, respectively, compared with 68 percent and 59 percent of LSMS+ households. While these coping strategies remained the most commonly used during the pandemic, an increasing share of households took out loans from a financial institution, delayed payments of financial obligations, or engaged in additional income-generating activities by December (Round 4). At the same time, there was a striking increase in the share of households that had received government assistance. Moreover, more households received assistance from NGOs although to a lesser extent compared to government assistance. Figure 12: Actions adopted by households in response to COVID-19 crisis LSMS+ sample IDPoor sample with valid equity card Reduce food consumption 65 Reduce food consumption 89 68 86 Reduce non-food consumption 61 Reduce non-food consumption 84 59 72 Credit purchases 36 Credit purchases 65 43 59 Borrow from friends & family 27 Borrow from friends & family 45 29 51 Assistance from government 5 Assistance from government 45 27 96 Delay payment obligations 20 Delay payment obligations 31 26 35 Additional income generation 17 Additional income generation 22 25 36 Loan from financial institution 13 Loan from financial institution 16 22 28 Sale of assets 17 Sale of assets 23 15 18 Sell harvest in advance 5 Sell harvest in advance 9 6 6 Assistance from NGO 9 Assistance from NGO 23 13 32 Children took on chores 14 Children took on chores 7 7 8 Assistance from friends &… 18 Assistance from friends &… 16 6 4 Take pay advance from… 24 Take pay advance from… 6 8 Rely on savings 20 Rely on savings 10 3 2 Covered by insurance policy 0 Covered by insurance policy 0 0 0 % % Round 1 Round 4 Round 1 Round 4 2 On June 24, 2020 Cambodia launched a nationwide cash relief program to support poor and vulnerable SOCIAL households during the COVID-19 pandemic. The program provides cash transfers to households identified by the ASSISTANCE government as part of the Identification of Poor Households Program, known as “IDPoor.” The program is a temporary measure intended to alleviate the economic hardships that the poor and vulnerable face because of COVID-19. The government initially intended to provide cash payments to households covered under the IDPoor program for 7 months (June–December 2020). As the pandemic continued, the Prime Minister announced that the relief program would be extended until March 2021. In June 2020, Prime Minister Hun Sen had announced that the program would spend US$25 million a month to support IDPoor households and would benefit around 560,000 families.2 In December, the government had spent US$30.7 million to support more than 680,000 households (2.7 million people).3 2 Associated Press. June 24, 2020. “Cambodia to provide cash to poor hit by COVID-19 outbreak.” 3 The Ministry of Social Affairs, Veterans and Youth Rehabilitation. Page 9 REPORT NO. 4 MARCH 2021 Since June 2020, there has been a marked increase in the share of eligible IDPoor households receiving social assistance from the government.4 The share of eligible IDPoor households that received any type of social support increased considerably from 50 percent to 91 percent between June and August 2020, then remained high at 92 percent in October before reaching 95 percent in December (Figure 13). Of the eligible IDPoor households receiving social assistance, nearly all received social assistance provided by the government and nearly all assistance in the form of direct cash transfers (Figure 14 and Figure 15). In terms of distribution of specific social assistance programs, 90 percent of programs listed by households are the COVID-19 relief cash transfers for poor and vulnerable households, while 30 percent are the conditional cash transfer for pregnant women and women with children under the age of 2. Figure 13: Percentage of Figure 14: Percentage of IDPoor Figure 15: Percentage of IDPoor social assistance beneficiaries households reporting main sources households reporting types of social among eligible IDPoor of social assistance assistance received households 91 92 95 95 94 % IDPoor households % IDPoor households 50 % 9 3 1 2 3 R1 R2 R3 R4 Government Food NGO Direct cash transfers Other in-kind transfers (excl. food) Most IDPoor households had received 6 to 7 installments of the COVID-19 relief cash transfers by December/January, averaging a total of US$252. Those who verified the validity of their equity card with the local authorities immediately following the launch of the program would have received about 7 installments around the time Round 4 of the HFPS was implemented.5 Among the IDPoor cash transfer beneficiary households, 24 percent had received 7 transfers, 37 percent 6 transfers, 20 percent 5 transfers, and 10 percent had received 4 transfers since the launch of the program in June 2020 (Figure 16). Only 9 percent report having received three or fewer transfers, likely due to later registration. The number of received installments can also vary as households enter or exit the IDPoor program over time. Since the launch of the COVID-19 relief cash transfer program, IDPoor households have received an average of US$252 (Figure 17). As expected, the average total cash transfer per household increases with the reported frequency of receipt. 4 Eligible IDPoor refers to an IDPoor household with a valid equity card. 5 Registration or validation of the equity card with the commune council is a prerequisite for receiving the relief cash transfers. Page 10 REPORT NO. 4 MARCH 2021 Figure 16: Number of times cash transfers were Figure 17: How much cash was received in total, by received number of times cash transfers were received 53 313 48 43 271 252 37 234 31 195 USD 24 148 % 20 16 104 10 49 422 4 6 1 Once Twice Three Four Five Six Seven Once Twice Three Four Five Six Seven Total times times times times times times times times times times Round 2 Round 3 Round 4 Round 4 The relief cash transfer program has reached many IDPoor households and registration has increased, and few eligible IDPoor households remain uncovered. Only 7 percent of households with an equity card had yet to receive the cash transfers in December (Figure 18). Most of these households (59 percent) had not registered at the commune council because they were not aware of the program or did not know that the cash transfers were for them. Registration among those that have not yet received transfers increased between August and October and then stagnated in December (Figure 19). While the reach of the program is high with 93 percent of eligible IDPoor households having received the transfers, it has not further expanded into universal reach since October. Figure 18: Share of eligible IDPoor households that Figure 19: Registration among eligible IDPoor have received the government relief cash transfers households that have yet to receive relief cash transfers 10 8 7 60 59 74 % % 90 92 93 40 41 26 R2 R3 R4 R2 R3 R4 Yes No Yes No The relief transfer program is perceived to have contributed to the economic well-being of beneficiary households. Around 37 percent of IDPoor households receiving the relief cash transfers report that the program has been extremely “important” for their household's economic well-being, while another 39 percent report that it has been “very important” (Figure 20). As a result, the relief program is perceived to have had an impact on the economic well-being of beneficiary households and made “a lot of difference” for 47 percent of these households (Figure 21). Beneficiaries of the COVID-19 cash transfers largely spent their cash on food (Figure 22). Page 11 REPORT NO. 4 MARCH 2021 Figure 20: How important was the relief transfer Figure 21: How much of a difference (on a scale from 1 for household economic well-being? to 5) did the relief transfer make to household economic well-being? 3% 24% 23% 37% 47% 39% 27% Extremely important Very important Moderately important Not so important 1 - no difference 2 3 4 5 - a lot of difference Not important at all Figure 22: How the COVID-19 relief cash transfers were used 100 Food 100 99 Other essential 58 54 items 64 15 Paid back loan 8 12 2 Other 5 5 0 20 40 % 60 80 100 Round 2 Round 3 Round 4 Food insecurity remained unchanged during October–December 2020 following a decline during August– October. Among IDPoor households in Round 4, the prevalence of moderate-or-severe food insecurity during the FOOD previous 30 days preceding the survey was 34 percent (± 3.64), including 0.9 percent (± 0.56) who were severely INSECURITY food insecure (Figure 23).6 This is compared to 39 percent (± 4.13) affected by moderate-to severe food insecurity and 2.5 percent (± 1.09) by severe food insecurity in Round 3, a prevalence of food insecurity statistically indistinguishable from Round 4. Food insecurity also remained unchanged among the surveyed LSMS+ households, albeit from a lower base. Among the LSMS+ sample, 17 percent (± 5.62) of the population had experienced moderate-or-severe food insecurity in the past 30 days in December (Round 4), of which 0.5 percent (±1.00) were affected by severe food insecurity. In October (Round 3), 17 percent (± 5.34) experienced moderate-or-severe food insecurity, of which 0.7 percent (± 1.07) experienced severe food insecurity. The COVID-19 relief cash transfer program, whose poor and vulnerable beneficiaries largely spent their cash on food, likely helped mitigate food insecurity. 6 Food insecurity prevalence rates are based on the application of the Food Insecurity Experience Scale (FIES), an experience-based metric of food insecurity severity that relies on people’s direct responses to eight questions about their access to food. Pe ople experiencing moderate food insecurity will typically eat low-quality diets and might have been forced to also reduce the quantity of food they would normally eat at times, while those experiencing severe levels would have gone for entire days without eating, due to lack of money or other resources to obtain food. See (http://www.fao.org/in-action/voices-of-the-hungry/sdgs/en/). Page 12 REPORT NO. 4 MARCH 2021 Figure 23: Share of population experiencing food insecurity 67 48 39 % 34 17 17 5 0.7 0.5 4 3 0.9 R2 R3 R4 R2 R3 R4 LSMS+ IDPoor Total population Moderate-or-severe food insecurity Severe food insecurity Page 13 REPORT NO. 4 MARCH 2021 BOX: SURVEY METHODOLOGY The Cambodia COVID-19 HFPS consists of two separate samples: (a) Living Standards Measurement Study Plus (LSMS+) and (b) IDPoor households. LSMS+ is a nationally representative household survey implemented October–December 2019 by the National Institute of Statistics (NIS) with technical and financial support from the World Bank. LSMS+ consists of 1,512 households, of which 1,364 have a phone number. The phone survey successfully reached and completed interviews for 700 of 1,364 households in May 2020, generating a response rate of 51 percent. In August 2020, 612 households were successfully re-interviewed. In October 2020, 469 households were successfully re-contacted from August, while 437 households were interviewed in all three rounds. In December 2020, 410 households were successfully interviewed, of which 368 households were re-interviewed from October and 342 households were interviewed in all four rounds. No replacement households have been included. Sampling weights were adjusted according to the steps outlined in Himelein (2014) to obtain unbiased nationally representative estimates from the sample.7 LSMS+ sample weights were computed to ensure representativeness at the national and urban/rural level. To monitor and evaluate the Cash Transfer Program for Poor and Vulnerable Households during COVID-19, which was launched in June 2020, an additional 1,000 IDPoor households were sampled. Before selection, the beneficiary list of the conditional cash transfer for pregnant women and women with children under the age of 2 was assigned into five strata: (i) Phnom Penh and other urban areas, (ii) Plain, (iii) Tonle Sap, (iv) Coastal, and (v) Plateau and Mountain. The sample was randomly selected proportional to the number of IDPoor households in each stratum. The phone survey successfully interviewed 984 households in June. In August, 784 households have been re-interviewed and 271 replacement households were added. Of these, 841 were successfully reached again in October, with 527 interviewed in all three rounds. In Round 3, 558 replacement households were also included. In December, 1,277 households were successfully interviewed, of which 945 households were re-contacted and 332 households were added as replacement households. A total of 675 households were interviewed in all four rounds. Sampling weights for the IDPoor sample were computed to ensure the representativeness of IDPoor households at the national and regional level. The HFPS was implemented using Computer Assisted Telephone Interview (CATI) techniques and the questionnaire was programmed using the Survey Solutions CAPI software package. Enumerators used mobile phone devices. Enumerators were given data bundles, allowing for internet connectivity and for daily data transfer and synchronization with the server. Field supervisors reviewed the survey responses with enumerators via one-on-one calls daily and addressed concerns that arose immediately following enumerators interview. At the same time, a research analyst was in charge of checking the uploaded data daily to identify errors so as to inform the field supervisors and enumerators. Data collection parameters, LSMS+ Round 4 ❖ Data collection period: 17 December 2020–12 January 2021 ❖ Completed interviews: 410 households (133 urban; 277 rural) ❖ Average duration of interview: 25 minutes Data collection parameters, IDPoor Round 4 ❖ Data collection period: 17 December 2020–12 January 2021 ❖ Completed interviews: 1,277 households (68 urban; 1,209 rural) ❖ Average duration of interview: 24 minutes 7 Himelein, Kristen (2014), Weight Calculations for Panel Surveys with Subsampling and Split-off Tracking, Statistics and Public Policy, 1:1, 40-45. Page 14