The Impact of the COVID-19 Crisis on Low Income Households in the Philippines: Deepening Distress despite Rebounding Economy1 COVID-19 Low Income HOPE Survey Note No. 2 Yoonyoung Cho, Doug Johnson, Yasuhiro Kawasoe, Jorge Avalos, and Ruth Rodriguez Executive Summary • The findings below are based on four rounds of Low Income Household Panel Economic (HOPE) surveys carried out between December 2019 and August 2020. The sample follows individuals in the country’s flagship social protection program, Pantawid Pamilyang Pilipino Program (Pantawid or 4Ps) beneficiary households, and other comparable low- income households. This is the second note of the series using the Low Income HOPE survey and it presents the welfare impact of COVID-19 on low-income households in the Philippines. • Overall employment recovered to 84 percent of its pre-crisis level by August 2020 as quarantine measures were gradually relaxed and economic activities resumed. Initial employment declines were largest in Luzon and in highly urbanized cities, and these places fell behind in recovery despite large rebounding in economic activities. • Households’ earnings remained much lower than pre-crisis levels. While employment has recovered significantly by August, total household earnings were still much lower than pre-crisis levels. Households had a reduced number of working adults and individuals worked fewer days in the labor market. • Household distress was deepening, and an increasing share of households have resorted to drawing down savings and reducing food and non-food consumption despite the rebounding economy. Levels of assistance have remained the same between June and August, but the extent to which households resort to coping mechanisms has increased. • An overall improvement in food security from June to August masks significant concerns among a small but non- negligible share of households facing a worsening condition. In particular, households in Luzon and in the lower end of the earnings distribution are more likely to report declining food security. • The COVID-19 crisis appears to have a significant impact on people’s mental health as a large portion of respondents reported poor mental health (e.g., depression) in August. Nearly a third of respondents cited suffering from severe depression; this indicator was strongly correlated with food insecurity. • A large share of low income households received government assistance through the social amelioration program (SAP). While many households in 4Ps were unaware that they received SAP, almost all Pantawid households received SAP by early April. Among non-4Ps households, close to 70 percent reported receiving at least one SAP payment and 16 percent reported receiving two SAP payments by August. Households with lower per capita household earnings prior to COVID-19 or those that experienced a greater earnings shock in April were more likely to be associated with SAP receipt. • Despite high coverage and prioritization of more vulnerable households, given the continuing and often deepening household distress, more targeted assistance can be considered going forward. Moreover, it is noted that SAP recipients spent, on average, nearly 3 hours receiving their first payment. Most of this time was spent waiting in line, highlighting the need for improving social protection delivery systems. 1 The note is prepared by the World Bank team led by Yoonyoung Cho, comprising of Doug Johnson, Yasuhiro Kawasoe, Jorge Avalos, and Ruth Rodriguez. The project builds on the impact evaluation implemented by the East Asia and Pacific Gender Innovation Lab (EAPGIL), and the Innovations for Poverty Action in the Philippines undertook data collecting activities. The research team at the Department of Social Welfare and Development provided valuable comments. The team thanks Eddy Trang for excellent research assistance. The work would not have been possible without generous financial support from the Department of Foreign Affairs and Trade (DFAT) of the government of Australia. 1 1. Introduction The COVID-19 pandemic and efforts to contain the virus have caused severe health and economic effects in the Philippines. As of mid-December in 2020, over 440,000 Filipinos have contracted COVID-19 and 8,677 have died from the virus.2 The Philippine economy contracted by 10.0 percent, year-on-year, in the first three quarters of 2020.3 In October 2020, the World Bank forecasted that the Philippine economy would contract by 6.9 percent in 2020, by far the largest drop for any Southeast Asian country.4 The COVID-19 pandemic threatens to reverse the trend of a steady decline in poverty in recent years, and to put close to 3 million additional Filipinos into poverty.5 The nationwide Labor Force Survey (LFS) surveys show that the labor market is gradually recovering from the peak of quarantine measures seen in April 2020, but still remains weak. Other nationwide studies and surveys, such as the World Bank’s High Frequency Monitoring Survey and Social Weather Station survey, have also highlighted significant impacts that the pandemic has brought to Filipinos’ lives and wellbeing. This study is the second of its series investigating the impact of COVID-19 and its policy measures on the outcomes of low income households using the Low Income Households Panel and Economic (HOPE) survey data. The first note (Cho et al. 2021) used Wave 0 (December 2019) and Wave 1 (April 2020) data to look into the labor market and food security outcomes, and examined the role of social protection programs. We found that the COVID-19 crisis had led to a severe drop in employment among sample households in April, with households in urban areas and those working in industry affected most. Food insecurity was also quite prominent; however, we found that the timely provision of emergency top-up cash assistance provided to beneficiaries of the country’s flagship Pantawid Pamilyang Pilipino Program (Pantawid or 4Ps) helped households cope with this challenge. The second note uses four rounds of the HOPE survey from December 2019 through August 2020, including both 4Ps and non-4Ps low income households, and monitors the impact of the COVID-19 crisis on poor and near-poor households in the Philippines. Wave 2 and Wave 3 surveys were conducted in June 2020 and August 2020, respectively. At the onset of the COVID-19 crisis, the Government of the Philippines (GoP) introduced strict lockdowns which evolved over time (Figure 1). Within a week of the World Health Organization (WHO)’s declaration of the pandemic on March 11, 2020, President Rodrigo Duterte declared the entirety of the Philippines to be under a State of Calamity for a period of six months and imposed strict enhanced community quarantines (ECQs). The ECQs were initially introduced in Metro Manila on March 15 and then expanded to the entire island of Luzon on the next day. By late March and lasting until May 16, almost all parts of the country were either under ECQs or Modified ECQs (MECQs) with a partial and limited relaxation of business operation. The Philippines’ ECQs, one of the most stringent measures in the region,6 restricted people’s movements except for essential purposes (related to medical and health conditions, for instance) and enforced the closure of nearly all non-essential shops and stores. With growing concerns on the economy as shown in the record high unemployment rate of 17.6 percent in April LFS, GoP started relaxing the quarantine rules. By June 2020, when the Wave 2 survey was conducted, all 17 regions in the country were under General Community Quarantines (GCQs) or Modified General Community Quarantine (MGCQ) except for a few locations such as the Cebu city. Under GCQ/MGCQs, more businesses and public transportation came back to operation albeit at a limited capacity. In August when the Wave 3 survey was conducted, however, as the number of COVID-19 cases rapidly increased, some parts of the country including 2 “Information on Epidemiology and Health System Capacity. Dashboard Developed in Collaboration with Jason Haw and Thinking Machines Data Science.” 3 World Bank 2020b. World Bank Philippine Economic Update: December 2020 edition. 4 World Bank 2020a. From Containment to Recovery: Economic Update for East Asia and the Pacific October 2020. 5 Measured against the lower middle-income poverty line of US$3.2/day 6 Lockdown stringency based on the COVID-19 Government Response Stringency Index by Oxford University. See Hale, Thomas, Sam Webster, Anna Petherick, Toby Phillips, and Beatriz Kira (2020). Oxford COVID-19 Government Response Tracker, Blavatnik School of Government. 2 Metro Manila went back to localized MECQs where local government authorities exercised the discretion to impose stricter quarantines. Figure 1. Community Quarantine Measures in the Philippines We find that, between April and August, employment rebounded substantially to about 84 percent of its pre- COVID-19 level. This is in line with the findings from April and July LFS (World Bank 2020b). Despite the recovery in employment, household earnings are still much lower than they were before the crisis. The total number of working adults per households was still less than before and individuals were working fewer days each week. Overall food security is improving, but concerns remain as vulnerable populations experience an increasing degree of food insecurity, and nearly a third of respondents in the sample reported suffering from symptoms of depression in August. A large share of households in our sample received SAP but its delivery met some implementation challenges. The remainder of the note is structured as follows. Section 2 provides background on the Pantawid program and the government’s immediate social protection response to the COVID-19 crisis. Section 3 details the data and methodology used for the study. Section 4 presents the results on the impact of the COVID-19 crisis on low-income households. Section 5 presents the delivery and performance of SAP. Finally, Section 6 summarizes and concludes the study. 2. Background and context The GoP introduced a large-scale Social Amelioration Program (SAP) to tamper the negative impact of COVID- 19 and its quarantine measures on households. In March 2020, the Bayanihan to Heal as One Act (Bayanihan 1) was enacted, which focuses on increasing funding for the health sector, safety nets for poor and vulnerable groups, and measures to financially support micro, small, and medium enterprises and jobs. About 18 million Filipino households were supported through the first tranche of transfers from the SAP, including top-up assistance for the beneficiaries of the country’s flagship 4Ps.7 Worth over US$4 billion (equivalent to 1.1 percent of the country’s Gross Domestic Product), the SAP emergency subsidies were one of the world’s largest cash assistance in terms of the proportion of population covered. The announced SAP intended to provide cash 7 See the Cho et al. 2021 for more details on the Pantawid program and the first tranche of SAP. 3 assistance of PhP 5,000 (US$100) to PhP 8,000 (US$160), roughly corresponding to a monthly wage of workers earning around the minimum wage and households’ subsistence expenditures in each region, up to two rounds of transfers. The initial implementation of SAP faced challenges in identifying new beneficiaries beyond Pantawid and transferring funds to them. Given that there was no ready-made list of poor and vulnerable populations available,8 beneficiaries had to be identified by a new application process. The Department of Social Welfare and Development (DSWD) brought in a paper application form and manual registration as commonly used during disaster relief efforts, and Local Government Units (LGUs) prioritized poor and vulnerable populations based on their local knowledge. However, the sheer volume of the target population was too large and operations amidst the pandemic faced several challenges. Also, unlike 4Ps beneficiaries who already had a digital channel (i.e., cash card issued by the Land Bank Philippines [LBP]) to receive funds from the government, new beneficiaries had to rely on LGUs for physical delivery of the cash assistance. These processes led to significant delays and confusion among non-4Ps beneficiaries receiving SAP. DSWD’s record shows that the transfer of the 1st tranche of SAP top- up benefits for 4Ps beneficiaries was completed by April 5, but that the distribution of SAP 1st tranche for non-4Ps households lasted until June 2020. The second tranche of SAP introduced a couple of changes. The target beneficiaries for second tranche was reduced to around 14 million households, including 5.3 million vulnerable “waitlisted” families who could not receive the first tranche in addition to 8.5 million vulnerable families living in ECQ areas as of May 1-15, 2020 (1.3 million Pantawid beneficiaries and 7.2 million other vulnerable families). On payment delivery, DSWD worked with six financial service providers (FSPs) including banks and mobile money issuers to provide beneficiaries with account opening services and to transfer subsidies through the accounts. The 2nd tranche SAP distribution took place from June to November 2020. As a result, with respect to overall SAP implementation, some households received two tranches while others received only one or none at all; further, some received funds through physical cash delivery while others received funds through digital channels. 3. Data We use data from a sample of 580 low income households, and 1,614 adult individuals from the sample households (See Cho et al. 2021 for more details). The sample took advantage of an already-constructed set of 4Ps and non-4Ps households that were used for previous impact evaluation studies of the 4Ps program.9 The data have a nationwide coverage of low income households with 4Ps and comparable non-4Ps households. The comparison between our sample and low income households in the nationally representative Family and Income Expenditure Survey (FIES) 201810 shows that our sample captures well the characteristics of poor and near-poor populations of the country. The sample households were initially interviewed prior to the COVID-19 shock face to face, and over the phone in subsequent waves (Table 1). The household roster was constructed, based on which individuals’ data for adults were collected. Response rates have been high without a significant difference between 4Ps and non-4Ps 8 The National household targeting system (Listahanan) covers around 70 percent of population with socio-economic status at the household level, but the information was outdated as data were collected in 2015. Enumeration for a new Listahanan was ongoing until the activity was suspended due to the pandemic in March 2020. 9 Given that 4Ps eligibility is determined based on a PMT score, the sample originally selected those 4Ps and non-4Ps with a similar level of PMT that were used for a regression discontinuity approach for impact evaluations. 10 To identify poor and near-poor households from the FIES, we restricted the FIES sample to households with per capita income between the 10th and 40th centile for the household’s respective region. 4 households.11 We present the results from the balanced panel data, based on the households who appear in all four rounds. Table 1. Sample Size by Wave Wave Survey Dates Sample size 4Ps Non-4Ps (households) 0 December 2019 580 292 288 1 April 2020 527 268 259 [91.8%] [89.9%] 2 June 2020 514 256 258 [99.2%] [99.6%] 3 August 2020 490 249 241 [97.3%] [93.4%] 4 October 2020 (not used for this 416 211 205 note) [84.7%] [85.1%] * Response rates from the previous round are in the brackets. Learning from earlier waves, in Waves 2 and 3, we strengthened key indicators to reflect the welfare of low income households. The survey includes detailed information on the households’ demographic characteristics and individuals’ labor market outcomes. Similar to previous waves, we investigate the experience of receiving social assistance. In addition, we constructed a more detailed index for food insecurity to further investigate the topic, and added other indicators such as coping mechanisms and mental health. Given that schools were still closed for June and August 2020, detailed modules on education and human capital are reserved for the following wave in October 2020. 4. Impacts of COVID-19 on low income households’ welfare Employment and Income The employment ratio dropped dramatically in April but recovered to 84 percent of its pre-crisis level by August (figure 2). The first note using the HOPE survey (Cho et al. 2021) found that the employment ratio, defined as the share of adults 18 and over who reported working at least one day in the previous 7 days, declined dramatically between December 2019 and April 2020. In December 2019, the employment ratio was 56 percent but dropped to 32 percent in April.12 In June, the employment ratio recovered to 46 percent and by August the employment ratio was 47 percent. Employment losses between December and April were particularly large in Luzon and in urban areas. Fortunately, workers in these areas benefited from strong rebounds in June and August. Nonetheless, in Luzon, the employment ratio in August 2020 is only 75 percent of that in December 2019 while it is 87 percent in Visayas and 93 percent in Mindanao. Likewise, employment recovery in urban areas lags behind rural areas, with the employment ratio in August being 79 percent in urban areas, in contrast to 87 percent in rural areas. 11 As the economy is slowly returning to activities, higher attrition is observed in October survey, the implications of which will be further examined in the next note. 12 Our results for the baseline and April rounds of the survey differ slightly from the earlier note as we only include households who were surveyed in all four rounds. However, the overall magnitude of impact remains consistent with the earlier note. 5 Figure 2. Employment Ratios by Location The size of the shock and recovery pace differ by the sector of employment (Figure 3 left). Employment was more volatile for the industry sector workers compared to those in agriculture or services. Operations in construction and manufacturing, which are part of the industry sector, are sensitive to the quarantine rules. When many regions were put in GCQs in June, the industry sector rebounded strongly, but in August when some localities went back to stricter MECQs, sector activities were dampened again. The labor market performance affect households’ abilities to earn income. Figure 3 (right) displays the distribution of households by the number of working adults. In the pre-COVID-19 economy, the share of households with no working adults is only 7 percent and almost 44 percent of sample households had two or more working adults. During the peak of ECQ in April, the share of households without any working adults soared to over 40 percent. In June and August, despite some employment rebounds, nearly a fifth of households still had no working adults, and only a third of households had two or more working adults. Figure 3. Level of Employment and Number of Those Working To date, unlike other countries, the pandemic has not shown large gendered impacts on the labor market with few changes in the gender makeup of the workforce. In other countries, the COVID-19 crisis has led to a large number of women, especially those in households with young children, dropping out of the workforce. 13 In December 2019 Wave 0 survey, 64 percent of working adults were men. In subsequent rounds, this proportion varied from 57 percent (in April) to 62 percent (in June). The proportion of the workforce made up of women in households with children ages 5 and under has actually increased since the onset of the crisis, from 13 percent at 13 https://time.com/5900583/women-workforce-economy-covid/ 6 Wave 0 to 17 percent in Wave 3 in August. We caution however that these surveys were conducted before the start of public school classes in October 2020. Employed adults worked fewer days in June and August than April but for modestly higher wages. Figure 4 (left) displays average days worked per week among working adults by round. While the employment ratio increased between April and June, average days worked fell significantly from 5.25 to 4.65, and stayed at the similar levels between June and August. The lower employment ratio, combined with fewer number of working adults and average working days, led to large declines in household earnings. Figure 4 (right) displays mean and median weekly household earnings per capita with no earnings recorded as zero. These figures highlight the depth of the COVID-19 shock in household earnings and employment. Moreover, they reveal that while the employment ratio has somewhat recovered, the combined impact on earnings and employment still continues. Figure 4. Mean Days Worked among Employed Adults and Weekly per Capita Earnings Note: Whiskers represent 95 percent confidence intervals. Coping actions and assistance Nearly all households were forced to engage in and continue with coping actions. The majority of households (94.4 percent in June and 97.5 percent in August) reported that they engaged in at least one coping action such as borrowing or reducing consumption since the start of the crisis. The most commonly used coping actions were reducing consumption, especially food, followed by borrowing from friends and drawing down savings (Table 2). These results on coping mechanisms are quite similar to the World Bank’s High Frequency Monitoring Household Survey conducted in August.14 The most common sources of loans were friends and family and sari-sari stores: 59 percent borrowed from friends or family and 40 percent borrowed from sari-sari stores. In most cases, the total amount borrowed was relatively small—among borrowing households, 30 percent borrowed less than one week of total Wave 0 household earnings and 68 percent borrowed less than four weeks of total Wave 0 household earnings. Table 2. Prevalence of Coping Actions and Assistance Action or form of assistance Wave 2 Wave 3 Difference (June) (August) (Wave 3-Wave 2) Coping action Draw down saving (bank, cash) 48% 63% 16%p*** 14World Bank High Frequency Monitoring Survey Round 1 Results. http://pubdocs.worldbank.org/en/546181605520156388/Results- from-the-Philippines-COVID-19-Households-Survey-conducted-in-August-2020.pdf 7 Take out a loan from friends, banks, cooperatives, or other 61% 67% 5%p** lenders (including buying items on credit from a shop). In-kind loan from friends or relatives (need to pay back in kind) 32% 42% 10%p*** Reduced investment in farm or other business 18% 24% 6%p** Sold asset 6% 8% 2%p** Reduced food consumption 65% 71% 7%p** Reduced non-food consumption 53% 68% 15%p*** Change jobs 21% 25% 4%p Assistance Got Social Amelioration Program (SAP) assistance 84%§ 49%§§ n/a Got assistance from relatives/friends within the country 22% 21% 0%p Got assistance from relatives/friends in other countries 9% 8% -1%p (remittance from overseas) Got assistance from an NGO 20% 18% -1%p Note: *** p< 0.001, ** p<0.05, * p<0.1. In June, households were asked if they had engaged in any of the following actions since the start of lockdowns. In August, households were asked if they had engaged in any of the following actions since the last time they spoke with the surveyor. Means include all observations in each round but to generate p-values we restricted observations to those households in both rounds in order to conduct paired t-tests. § Yes if respondents received SAP assistance since lockdown or was a 4Ps beneficiary; §§ Yes if the respondent received SAP assistance since the last survey in June or was a 4Ps beneficiary. Despite rebounding employment, with continued labor market struggles, it appears that households’ distress is deepening. The prevalence of most coping actions increased significantly while assistance remains around the same levels between the Waves 2 and 3. In particular, the share of those drawing down savings and reducing non- food consumption has greatly increased by 30 percent over the two months. Taking in-kind loans also significantly increased. The likelihood of selling assets or reducing food and non-food consumption was far greater among households in the bottom 40 percent of the earnings distribution (based on per capita household earnings of Wave 0 in December 2019) compared to their better off peers. This indicates that low income households have relatively limited options for coping mechanisms. Food security With reduction of food consumption still one of the most common coping strategies, many households continued to face food insecurity in both June and August. The Wave 1 survey in April 2020 found about 56 percent of households reported that a family member ate fewer meals in a day because of lack of food in the previous 7 days. In the June and August surveys, we included additional questions adopted from the Household Food Security Access Scale (HFIAS) to more robustly gauge food security (see box 1 below for more details). Food insecurity is measured in the three dimensions – anxiety (Q1), inadequate quality of food (Q2-Q5), and insufficient food intake (Q6-Q9) – over the past 4 weeks. The anxiety and inadequate quality of food that is suboptimal against the respondent’s preferred food selection or diversity, reflect a subjective assessment of the adequacy of food security while food intake reflects the availability of food and incidence of hunger. We constructed a food insecurity index ranging from 1 to 5 with the higher value indicating the more food insecure (the details of food security index are discussed in the appendix).15 15Measuring food insecurity in a comparable and rigorous manner is still an important research issue. See a blog: https://blogs.worldbank.org/opendata/how-should-we-measure-food-security-during-crises-case-nigeria. While the index constructed here is not comparable with other studies or data, it enables the comparison among different groups within this survey over time. 8 Box 1: Construction of the food security index In June and August rounds of the survey (Waves 2 and 3), we included questions from the Household Food Insecurity Access Scale (HFIAS), commonly used to assess food security. The HFIAS is an experiential measure of food security, that is based on questions about households’ experience regarding food security. Experiential measures of food security such as the HFIAS or the closely related Food Insecurity Experience Scale differ from food intake measures such as the Food Consumption Score which gauge food security by asking more detailed questions about consumption of various foodstuffs. The nine questions which we used to gauge food security are listed below. These questions, and the method used to gauge food security differ slightly from the approach described in Coates et al. (2007). The original HFIAS questionnaire includes two stage questions for each action: a first question which asks whether the household has engaged in the action in the previous four weeks and, if the respondent answers “yes” to the first que stion, a second question about how often the household engaged in the behavior. In our survey, given that it is a phone interview within a limited time frame, we ask a single question while allowing “never” as part of the answer. The questions used are as follows, reflecting the three important dimensions of food insecurity —anxiety, food quality, and food intake. The respondents report the incidence with “never,” “rarely (1-2 times a month),” “sometimes (3-10 times a month),” and “often (more than 10 times a month).” • Q1 (Anxiety): In the last four weeks, how often were you worried that your family might not have enough to eat? • Q2 (Food quality): In the last four weeks, how often were you or any household member not able to eat the kinds of foods you preferred because of a lack of resources? • Q3 (Food quality): In the last four weeks, how often did you or any household member have to eat a limited variety of foods due to a lack of resources? • Q4 (Food quality): In the last four weeks, how often did you or any household member have to eat some foods that you really did not want to eat because of a lack of resources to obtain other types of food? • Q5 (Food quality): In the last four weeks, how often did you or any household member have to eat a smaller meal than you felt you needed because there was not enough food? • Q6 (Food intake): In the last four weeks, how often did you or any other household member have to eat fewer meals in a day because there was not enough food? • Q7 (Food intake): In the last four weeks, how often was there no food to eat of any kind in your household because of lack of resources to get food? • Q8 (Food intake): In the last four weeks, how often did you or any household member go to sleep at night hungry because there was not enough food? • Q9 (Food intake): In the last four weeks, how often did you or any household member go a whole day and night without eating anything because there was not enough food? Given the changes introduced to the way the questions were asked and differences in the reference time period, it is not feasible to directly compare with other studies that used the HFIAS to assess food security. However, this provides detailed information to gauge food insecurity within this sample over time. While the overall extent to which households experience food insecurity declined over time, there are still signs of serious concern. Those who experience hunger (Q8 and Q9) sometimes and often, whose food insecurity index is 5, increased from June to August even though the majority did not experience hunger at all and their share has increased (Figure 5). The share of those who had to compromise the quality of food has significantly declined 9 between June and August, indicating the recovery with respect to food security. Despite such progress, the share of households expressing the anxiety over food security has remained high. Figure 5. Change of Severity of Food Insecurity between June and August 2020 Food security improved in Visayas and Mindanao compared to Luzon, and among relatively better off households (Figure 6). While the majority of households reported no change or improvement in food security, some experienced a worsening of the situation. In particular, the share of households reporting improvement (i.e., decreases in food insecurity index) was significantly lower in Luzon and among households in the bottom 40 percent of household earnings distribution. 16 This indicates that the overall trend and average may mask significant vulnerability of the poorest households. While the overall situation is getting better, continued restrictions in economic activities especially in Luzon may have disproportionately large impacts on the poorest and most vulnerable households. Figure 6. Change in Food Security between June and August 2020 by Island Group and Earnings Level Mental health Overall mental health is poor, with many respondents displaying symptoms of severe depression. Anecdotal evidence and some survey data suggest that the COVID-19 crisis has led to a sharp decline in Filipinos’ mental health. According to news reports, suicides have increased drastically since the start of lockdowns 17 and a 16 Note that this is bottom 40 percent in our sample households which is already a sample of low income households in the country equivalent to bottom 40 percent of the overall income distribution. 17 “Church Asked to Help Curb Rising Philippine Suicide Rate - UCA News.” 10 nationwide phone survey conducted in June 2020 found that 30 percent of households reported that a member of their household had developed mental health symptoms.18 To assess the impact of the crisis on mental health, in August we asked respondents a question about their overall life satisfaction and five questions from the Mental Health Inventory 5 (MHI 5) question index.19 Nearly a third exhibited symptoms of severe depression (Figure 7 left). These values are far higher than typically observed in large populations. 20 MHI5 scores are strongly associated with food security, indicating that food insecure households tend to have significant mental health concerns (Figure 7 right). Correlations between MHI5 and earnings levels or employment changes are relatively weak compared to the visible association between the mental health measure and food security index. Figure 7. Share of Households by Mental Health Status and Mental Health Index by Food Insecurity Status 5. Delivery of SAP Implementation efficiency and quality matters in providing timely support for beneficiaries. Cho et al. (2021) using Waves 0 and 1 survey data highlighted the importance of a timely delivery of cash assistance during the time of crisis in mitigating its negative impact on households’ welfare. In a similar manner, we examined the delivery aspects of SAP and its targeting performance. A large share of 4Ps households did not seem to be aware that they received SAP or that DSWD has modified the regular program to respond to the COVID-19 pandemic. Most of 4Ps households received the 1st tranche of SAP top up benefits before April 5 and 1.3 million eligible 4Ps beneficiaries who were still in ECQ received the 2nd tranche by June 10. However, only 41 percent of 4Ps households were aware that they were receiving SAP despite information campaign efforts. Since SAP was delivered as a top-up to household’s existing 4Ps benefits and 4Ps’ education grants during school break were not provided, 4Ps households may have not realized that they were not receiving education grants, but instead were getting SAP top-ups. Similarly, only about 40 percent of 4Ps beneficiaries were aware that DSWD has modified their regular program such as family development sessions to be conducted online through Facebook; even among those who knew about it, only 18 percent visited the site. This suggests that there is a need to enhance awareness among 4Ps beneficiaries and to encourage their active participation in modified formats of program delivery. 18 Warren, Parkerson, and Collins, “RECOVR Philippines: Tracking the Effects of the COVID -19 Pandemic.” 19 We use cut-points from Yamazaki et al to determine depressive categories. 20 Yamazaki, Fukuhara, and Green, “Usefulness of Five-Item and Three-Item Mental Health Inventories to Screen for Depressive Symptoms in the General Population of Japan”; Houghton et al., “Tertiary Level Students and the Mental Health Index (MHI -5) in Ireland.” 11 Among non-4Ps low income households, close to 70 percent reported receiving at least one SAP payment and 16 percent reported receiving two SAP payments by August. It appears that more vulnerable households were slightly more likely to receive SAP. For instance, pre-COVID-19 per capita household earnings were lower for SAP households (PhP 405 per week) than non-SAP households (PhP 560 per week).21 Similarly, households with a greater earnings shock due to the pandemic seemed to be more likely to benefit from SAP—per capita household earnings decreased by 60 percent for SAP recipients between Waves 0 and 1 (from December 2019 to April 2020) whereas the decrease was about 40 percent for non-SAP households. The high coverage of SAP and prioritization of poor and vulnerable households are in line with the findings from the World Bank’s High Frequency Monitoring Survey. Nonetheless, given the size and duration of the shock, a more targeted approach to sufficiently address acute distress may be required. Most non-4Ps recipients of SAP received their Figure 8. Date of First SAP Payment for non-4Ps 1st SAP payment in cash at the LGU office 22 Households whereas nearly all 4Ps households received their benefits through LBP cash card. Most non- 4Ps recipients of SAP first tranche received their payment between late-April and mid-May (Figure 8). The date of first SAP payment did not vary much by island group or rural/urban status. Compared to 4Ps beneficiaries, SAP beneficiaries spent less time and money travelling to receive their cash. This is because there are only few LBP ATMs available for 4Ps beneficiaries (see Cho et al. 2021 for detailed discussions) whereas LGU offices are located in the community near beneficiaries. While rural 4Ps households traveled further to receive their payments compared to their urban counterparts, rural SAP households spent about the same amount of time traveling to receive their SAP payment as their urban peers. With respect to waiting time, non-4Ps SAP beneficiaries had to spend an average of 170 minutes at LGU offices to receive payments whereas 4Ps beneficiaries had to wait for average of 100 minutes at LBP ATMs. Overall, both 4Ps and non-4Ps households spend a significant amount of time and money just collecting their payments (Table 3). Table 3. Monetary and Time Costs to Receive Benefits Average cost Average travel Average wait time Average total time (one way, PhP) time (one way, (mins) (mins) mins) 4Ps pre-COVID-19 42.4 37.4 102.2 177.0 4Ps post-COVID-19 50.4 30.9 100.6 162.4 Non 4Ps SAP 13.4 14.4 170.0 198.8 21Due to the low awareness of SAP among 4Ps households, this analysis excludes 4Ps households. 22About 70 percent of recipients of the 2nd tranche of SAP received from local LGU office and 25 percent received from direct deposit through bank account. However, the sample size is too small to investigate the delivery efficiency of the 2 nd tranche. 12 6. Summary and Conclusions The COVID-19 pandemic and efforts to contain its spread led to a significant disruption in economic activities, and its lingering impact remains even if the economy is slowly rebounding. World Bank (2020b) reported that the labor market conditions gradually improved but that the recovery is fragile. July LFS provides a positive picture of the recovery of the overall labor force participation rate to the pre-pandemic level and reduced unemployment rate compared to the peak of quarantine in April. However, the wellbeing of low income households is far from being recovered. Our survey shows that in August, the fifth month since the quarantine measures, a significant share of households still remain without any labor income after the breadwinner lost his or her job. Continued restrictions in economic activities in Luzon, the economic powerhouse of the country, have had a lasting impact on the economy. More households draw down savings, reduce consumption, sell assets, and borrow from friends and other informal channels. Food insecurity reflected in household’s anxiety, adequacy of food diversity, and food intake, remains a grave concern for low income households. Amidst overall improvement, there are households which are experiencing a worsening condition. In particular, overall experience of hunger due to the lack of resources is no longer prevalent, but for a small share of households, the frequency of inadequate food intake appears to be increasing. Households’ confidence on food security has not improved much, which may have significant implications on their members’ mental wellbeing. The government provided a large-scale cash assistance despite many implementation challenges. As the first note using the COVID-19 Low Income HOPE Survey (Cho et al. 2021) discussed in detail, having a well-established social protection program with a solid delivery system was useful. SAP benefitted a large number of low income households, and we also found that poorer and more vulnerable households were more likely to receive SAP payments. However, we noted that provision of SAP benefits for non-4Ps came with a significant delay and due to weak delivery systems, many beneficiaries had to spend long hours waiting in line during the pandemic to receive benefits. Given the magnitude and duration of the COVID-19 impact, and continuing distress and food insecurity among pockets of the population in the midst of overall recovery, a more targeted intervention may be required. Further, efforts to strengthen delivery systems to enhance beneficiaries’ experience to utilize social assistance programs should be prioritized. 13 Reference Brown, Caitlin, Martin Ravallion, and Dominique van de Walle. 2020. “Can the World’s Poor Protect Themselves from the New Coronavirus?” Cambridge, MA: National Bureau of Economic Research. https://doi.org/10.3386/w27200. Brussevich, Mariya, Era Dabla-Norris, and Salma Khalid. 2020. “Who Will Bear the Brunt of Lockdown Policies? Evidence from Tele-Workability Measures Across Countries.” International Monetary Fund. IMF Working Paper, 20/88. Chetty, Raj, John N Friedman, Nathaniel Hendren, and Michael Stepner. 2020. “The Economic Impacts of COVID- 19: Evidence from a New Public Database Built Using Private Sector Data,” National Bureau of Economic Research Working Paper No. 27431. Coates, Jennifer, Anne Swindale, and Paula Bilinsky. 2007. “Household Food Insecurity Access Scale (HFIAS) for Measurement of Food Access: Indicator Guide: Version 3: (576842013-001).” American Psychological Association. https://doi.org/10.1037/e576842013-001. Cho, Yoonyoung, Jorge Avalos, Yasuhiro Kawasoe, Doug Johnson, and Ruth Rodriguez. 2021. "Mitigating the Impact of COVID-19 on the Welfare of Low Income Households in the Philippines: The Role of Social Protection." COVID-19 Low Income HOPE Survey Note No. 1, World Bank. FAO, IFAD, UNICEF, WFP, and WHO. 2020. The State of Food Security and Nutrition in the World 2020. FAO, IFAD, UNICEF, WFP and WHO. https://doi.org/10.4060/ca9692en. Houghton, Frank, Noreen Keane, Niamh Murphy, Sharon Houghton, and Claire Dunne. 2010. “Tertiary Level Students and the Mental Health Index (MHI-5) in Ireland,” Irish Journal of Applied Social Studies. Vol 10 (1). Iacobucci, Gareth. 2020. “Covid-19: Increased Risk among Ethnic Minorities Is Largely Due to Poverty and Social Disparities, Review Finds.” BMJ. https://doi.org/10.1136/bmj.m4099. Tableau Software. “Information on Epidemiology and Health System Capacity. Dashboard Developed in Collaboration with Jason Haw and Thinking Machines Data Science.” Accessed November 10, 202 0. https://public.tableau.com/views/COVID- 19CasesandDeathsinthePhilippines_15866705872710/Home?:embed=y&:showVizHome=no&:host_url= https%3A%2F%2Fpublic.tableau.com%2F&:embed_code_version=3&:tabs=no&:toolbar=yes&:animate_ transition=yes&:display_static_image=no&:display_spinner=no&:display_overlay=yes&:display_count=y es&publish=yes&:loadOrderID=0. ucanews.com. “Church Asked to Help Curb Rising Philippine Suicide Rate - UCA News.” Accessed November 11, 2020. https://www.ucanews.com/news/church-asked-to-help-curb-rising-philippine-suicide-rate/89264. Villanueva, Joann. 2020. “IMF Cuts 2020 Growth Forecast for PH; Hikes 2021 Projection.” Accessed November 10, https://www.pna.gov.ph/articles/1118440. Warren, Shana, Doug Parkerson, and Elliott Collins. 2020. “RECOVR Philippines: Tracking the Effects of the COVID- 19 Pandemic.” https://www.poverty-action.org/recovr-study/recovr-philippines-tracking-effects-covid- 19-pandemic. 14 World Bank 2020a. From Containment to Recovery: Economic Update for East Asia and the Pacific October 2020. https://www.worldbank.org/en/region/eap/publication/east-asia-pacific-economic-update. World Bank 2020b. Philippines Economic Update: Building a Resilient Recovery. World Bank December 2020. https://documents.worldbank.org/en/publication/documents- reports/documentdetail/983051607354214738/philippines-economic-update-building-a-resilient- recovery World Bank 2020c. Monitoring COVID-19 Impacts on Families and Firms in the Philippines. https://www.worldbank.org/en/country/philippines/brief/monitoring-covid-19-impacts-on-firms-and- families-in-the-philippines Yamazaki, Shin, Shunichi Fukuhara, and Joseph Green. 2005. “Usefulness of Five-Item and Three-Item Mental Health Inventories to Screen for Depressive Symptoms in the General Population of Japan.” Health and Quality of Life Outcomes 3, no. 1: 48. 15 Appendix 1: Construction of Food Security Index The food security index used in this analysis is similar to the Household Food Insecurity Access Scale (HFIAS) described in Coates et al. (2007) but includes one slight change. The HFIAS is based on a set of questions which asks about 9 food-consumption actions. For each food consumption action, the HFIAS includes two questions: a first question which asks whether the household has engaged in the action in the previous four weeks and, if the respondent answers “yes” to the first question, a second question about how often the household engaged in the behavior. In the answer of time, our survey only included a single question which asked about the frequency of engaging in each action but allowed respondents to respond “never” to the single question. While dropping the preliminary question saved time, it may have biased upwards our estimates of food security as respondents may be more likely to respond “no” when asked the first preliminary question “have you engaged in this action in the past four weeks?” than to respond “never” when asked “how often have you engaged in this action in the past four weeks?” This bias is of particular concern for the last three actions in the index. These actions are deemed so concerning that even if a respondent only rarely engages in these actions, the respondent is labeled severely food insecure. To correct for the potential bias caused by dropping the preliminary question, we also adjust the rule for determining food security status which labels households severely food insecure only if they respond “sometime” or “often” to the final three questions in the index. The following table describes our modified coding scheme to determine the food security status. Note that the final food security category for a household corresponds to the most severe food insecurity status for any question. Table 3. Modified HFIAS Coding for Food Security Index Question Variable name Rarely Sometimes Often In the last four weeks, how often were you worried that your family might not have enough Q1 to eat? In the last four weeks, how often were you or any household member not able to eat the kinds of Q2 foods you preferred because of a lack of resources? In the last four weeks, how often did you or any household member have to eat a limited variety Q3 of foods due to a lack of resources? In the last four weeks, how often did you or any household member have to eat some foods that Q4 you really did not want to eat because of a lack of resources to obtain other types of food? In the last four weeks, how often did you or any household member have to eat a smaller meal Q5 than you felt you needed because there was not enough food? 16 In the last four weeks, how often did you or any other household member have to eat fewer Q6 meals in a day because there was not enough food? In the last four weeks, how often was there no food to eat of any kind in your household because Q7 of lack of resources to get food? In the last four weeks, how often did you or any household member go to sleep at night hungry Q8 because there was not enough food? In the last four weeks, how often did you or any household member go a whole day and night Q9 without eating anything because there was not enough food? 1 2 3 4 5 Most secure Most insecure 17