SOLOMON ISLANDS HIGH FREQUENCY PHONE SURVEY ON COVID-19: RESULTS FROM ROUND TWO Data Collection: December 2020 – January 2021, and April 2021 Acknowledgement: This report was written by Kristen Himelein, Christopher Hoy, Darian Naidoo, Donny Pasaribu (Consultant), James Carroll Waldersee, and Shuwen Zheng (Consultant) of the World Bank using the first two rounds of the World Bank’s High Frequency Mobile Phone Survey (HFPS) to Assess the Socio-Economic Impacts of COVID-19 in Solomon Islands, and the first round of the UNICEF’s Social-Economic Impact Assessment Survey (SIAS) in Solomon Islands. The team gratefully acknowledges the Korea Trust Fund for Economic and Peace-Building Transitions for financing the data collection and analysis for the World Bank survey and the UNICEF Pacific Multi Country Office for financing the data collection and analysis of the UNICEF SIAS. The team also appreciates the contribution of Masoomeh Khandan in assisting with the analysis. The authors thank UNICEF colleagues Ronesh Prasad, Jun Fan, Zelalem Taffesse, Peta-Gaye Bookall, Chander Badloe, and Jayachandran Vasudevan as well as World Bank colleagues Virginia Horscroft, Lodewijk Smets, Giorgia Demarchi, Son Thanh Vo, Annette Leith, Kara Mouyis, Rinku Murgai, Shohei Nakamura, and Jason Sze for their comments on earlier drafts. The team also thanks C. Jason Smith for copyediting. The findings, interpretations, and conclusions expressed in this report are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development, the World Bank, UNICEF and its affiliated organizations, nor those of the Executive Directors of the World Bank or the governments they represent. SOLOMON ISLANDS HIGH FREQUENCY PHONE SURVEY ON COVID-19: RESULTS FROM ROUND TWO Data Collection: December 2020 – January 2021, and April 2021 TABLE OF CONTENTS Executive Summary........................................................................................... 4 1. Update on the COVID-19 situation in Solomon Islands............. 10 2. Data Collection......................................................................................13 2.1. Description of Survey Objectives & Instrument...............................13 2.2. Methodology and Data Collection.....................................................14 2.3. Re-weighting...........................................................................................15 2.4. Analytical approach................................................................................12 3. Employment and Income....................................................................17 3.1. Employment............................................................................................17 3.2. Income......................................................................................................21 3.3. Non-Farm Business............................................................................... 23 3.4. Agriculture............................................................................................... 24 3.5. Remittances............................................................................................ 27 3.6. Access to Financial Services................................................................ 28 3.7. Financial Anxiety.................................................................................... 30 4. Household Coping Strategies and Personal Coping Strategies..................................................................................31 4.1. Coping Strategies of Households.......................................................31 4.2. COVID19 Related Concerns and Coping Strategies of Adults and Children.......................................................................... 34 5. Government response......................................................................... 38 5.1 ESP Background..................................................................................... 38 5.2 Awareness of the ESP........................................................................... 38 5.4 ESP Grant Applications and ESP Subsidies....................................... 39 5.6 Financial Sector Initiatives....................................................................40 5.7 Payouts from National Provident Fund............................................. 41 6. Public Services....................................................................................... 43 6.1. Access to Health Care.......................................................................... 43 6.2. COVID-19 Vaccination......................................................................... 46 6.3. Water and Sanitation............................................................................. 47 6.4. Education and Schooling..................................................................... 49 2 7. Public Trust & Community Security................................................. 53 7.1. Public Trust.............................................................................................. 52 7.2. Community Security Issues................................................................. 54 8. Conclusions and Policy Considerations......................................... 58 8.1. Conclusions............................................................................................ 58 8.2. Policy considerations............................................................................60 Appendix 1: Technical..................................................................................... 62 A1.1. HFPS Instrument Design........................................................................ 62 A1.2. Sampling.................................................................................................... 62 A1.3. Weighting.................................................................................................. 62 Appendix 2: Tables........................................................................................... 63 Abbreviations COVID-19 Coronavirus Disease COVAX COVID-19 Vaccines Global Access DBSI Development Bank of Solomon Islands DHS Demographic and Health Survey ESP Economic Stimulus Package GDP Gross Domestic Products HFPS High-Frequency Phone Survey HIES Household Income and Expenditure Survey IMF International Monetary Fund MRP Multilevel Regression and Poststratification NGO Non-Governmental Organization PIC Pacific Island Countries SIAS The UNICEF Social-Economic Impact Assessment Survey UNICEF United Nations Children’s Fund USD United States Dollars WASH Water, Sanitation and Hygiene WHO World Health Organization 3 Executive Summary This report presents the findings of the second of five planned rounds of mobile phone surveys, as well as the results of a companion survey funded by UNICEF. The first World Bank High Frequency Phone Survey (HFPS) in Solomon Islands on the socioeconomic implications of the coronavirus pandemic found wide ranging impacts that deserve to be monitored as long as the pandemic continues. The second round of the HFPS interviewed 2,882 households across the country in December 2020 and early January 2021, on topics including employment and income, the economic stimulus package, coping strategies, public services, and public trust and security. The 2020-2021 HFPS survey was complemented by the UNICEF Social-Economic Impact Assessment Survey (SIAS) which collected data in April 2021 on 1,530 households that were re-interviewed based on recontact of all Round Two HFPS households. The SIAS covered areas of impacts including access to health care, family arrangements, education, coping strategies, and water and sanitation. The findings in this report supplement the previous HFPS findings and other data on macroeconomic conditions, firm-level information, and etc., to monitor the evolving impacts of the pandemic situation. While the findings are not without their caveats due to the limitations of mobile phone surveys, they provide an insightful picture of the overall impact on households of the first year of the COVID-19 pandemic. Partial employment recovery is evidenced across the second half of 2020, driven by recovery in formal sector jobs and urban areas. By December 2020, employment remained well below pre-pandemic levels. Round One results found there to be net loss in employment of between 7 percent and 11 percent of the pre-crisis workforce by June 2020. Based on Round Two data, 58 percent of adults reported working in January 2020 and 45 percent reported working in June 2020, confirming the fall in employment. In the Round Two survey of December 2020, 47 percent reported working, indicating a slight employment recovery in the second half of 2020, but this data was not statistically significant. The timing of the Round Two survey may mean that employment recovery is underestimated, as the main reason cited by respondent for stopping work between June and December 2020 was vacation or holiday 4 Executive Summary closures. For those living in Honiara, and other urban areas, around a quarter of the initial decrease seen between January and June had recovered by December, with weaker recovery in rural areas. Employment in the service sector faced a greater net decline than employment in agriculture from January to December 2020. While both the informal and the formal sector saw declines in the first half of 2020, the formal sector returned to 95 percent of January 2020 levels by December 2020, a stronger recovery than for the informal sector. The bottom 40 percent, third quintile, and the top 20 percent experienced no significant employment recovery from June to December, with only the fourth quintile experiencing significant recovery. The income impacts of the pandemic have been more widespread than the loss in employment and there is little evidence of recovery post-June 2020. The results also suggest that the income situation of rural areas is deteriorating. In Round Two (December 2020), relative to their “usual” income levels1, 12 percent of respondents reported increased income in the previous week, 43 percent reported the same level of income and 45 percent reported lower income or no income in the previous week. These percentages were not significantly different from Round One (June 2020) and may indicate negative welfare impacts of the pandemic situation for a large share of households at end of 2020. In rural areas, respondents were more likely to report lower than usual income in the past week in December 2020 (39 percent) than in June 2020 (30 percent). In contrast, urban areas did not see a more negative situation in relation to income in Round Two than in Round One. The share of households reporting less than usual income from non-farm enterprises was 51 percent in Round Two compared to 32 percent in Round One, a significant deterioration from June 2020 to December 2020. The apparent deterioration in non-farm income was most notable for rural households, with 48 percent reporting lower non-business income in the last week compared to 28 percent in Round One. While most households earning an income from agriculture could perform agricultural activities normally, 45 percent expected lower income from agricultural activities compared to the previous growing season. Overall, the picture in relation to income is not indicative of recovery and suggests rural areas are less income secure than urban areas. More than three-quarters of respondents (77 percent) felt worried about their household’s finances over the next month following the survey, with around a third being “very worried”. Similarly, respondents were pessimistic about the economy, with two-thirds (67 percent) expecting that the state of the economy would get worse in the next year. Econometric analysis indicates that respondents in households with a female-head of household were significantly more likely to have a negative outlook on the economic situation. For the minority of households that receive international Income at the individual level collected across Round One and Round Two. 1 5 remittances (3 percent), most remittances from international sources have declined or stopped in the past month relative to what they usually receive. Domestic remittances were more likely to have stayed at the same level. In terms of practical access to finance, of the minority or respondents who needed to go to a money agent in the month preceding the survey (12 percent), the majority (81 percent) were able to get access. The majority of households continue to use a mix of coping strategies, including reducing consumption, though there appears to have been a shift toward using strategies that could increase vulnerability in the long term. Half of all households (50 percent) are purchasing items on credit, with over a third borrowing from friends or family, leading to a potentially less secure financial future with relatively more households using strategies that increased debt than in Round One. Of particular concern is that 60 percent of households spent from savings. These coping strategies may degrade future productive capacity and resilience, even while most households continue to reduce consumption. Compared across survey rounds, households have relied less on informal safety nets in the second half of 2020, which may indicate the limited capacity of informal safety nets for widespread and ongoing crises situations. There was also a three-fold increase, from 5 percent to 17 percent, in the share of households reporting they had reduced the number of children attending school in their household, which may have human capital implications over the long term. In the UNICEF SIAS, around three-quarters of respondents from households with children reported that they were worried about COVID-19 and significant numbers reported concerns of behavioral impacts on children. Most respondent who had children in their household were worried about the potential for household members to become ill with COVID-19. The majority of these respondents were concerned about their parents getting sick (56 percent), and an even greater proportion were concerned about children getting sick (78 percent). The most common coping strategies employed by respondents in households with children were prayer (74 percent), staying connected with family/friends (42 percent), staying positive (32 percent), and reading the news (27 percent). Prayer and reading the news were far more common strategies in male-headed households, while staying connected with family and friends and staying positive were far more common coping strategies in female-headed households and those without children. In the UNICEF SIAS, around half of the respondents with children in their households, noticed a change in the behavior of children since the start of the COVID-19 pandemic. The most reported change in behavior of children was that children were worried (particularly that their family/friends would die or get sick from COVID-19). Other commonly reported changes in behavior from parents included children not completing homework, being more “clingy” than usual, and being angry. 6 Executive Summary Awareness of the Economic Stimulus Package (ESP) was uneven by December 2020. The SIG’s Economic Stimulus Package (ESP) was launched in May 2020 to support households and businesses during COVID-19. Around 60 percent of households had heard of the ESP at the time of the Round Two HFPS. Respondents in rural areas were less likely to have heard of the ESP (54 percent), compared to Honiara (81 percent). 83 percent of the top 20 percent of households had heard of the ESP compared to only 44 percent of the bottom 40 percent of households. Of households involved in the agricultural, fishing, and forestry sector, around a quarter had reportedly applied to the ESP to receive grants. Households in the top 40 percent, in urban areas and with a male-head, were more likely to have had applied for grants. Of households who had heard of the ESP and had a loan with a formal lending institution, 8.6 percent received financial relief through delaying loan repayments. The next round of the HFPS will examine issues of take up rates of the ESP elements, relative to eligibility. Of those households needing urgent care, routine care, or preventive care in the month prior to the World Bank HFPS survey, the vast majority had access. Of those who needed urgent care, 93 percent had access; of those needing routine care, 88 percent had access; and of those needing preventative care, 96 percent had access. Given the absence of COVID-19 cases and absence of mobility restrictions, high levels of health access are not surprising. However, based on results from the UNICEF SIAS, almost one fifth of respondents needed to buy medicine in the week prior to the survey and of these, 43 percent experienced difficulties in access. Female-headed households were more likely to have had difficulties (82 percent), compared to male-headed households (35 percent). COVID-19 vaccine hesitancy was widespread and significant at the time of the UNICEF SIAS. Around half of respondents (47 percent) responded that they would get vaccinated if they were offered an approved vaccine at no cost, while almost a third (29 percent) responded that they would not, and around one quarter (24 percent) were unsure. The most common concern was about side-effects (30 percent). The two other common concerns were mistrust in vaccines in general, and safety of the COVID vaccine in particular. Future rounds of the World Bank HFPS will probe vaccine hesitancy further, allowing for deeper analysis. According to the UNICEF SIAS data, around three-quarters (73 percent) of households had adequate access to water and soap for hand-washing. Around one fifth (22 percent) of households reported having insufficient drinking water in the last week. The main reasons households were unable to access drinking water were due to inadequate supply (54 percent) and a lack of affordability (22 percent). Of UNICEF SIAS respondents, half (50 percent) reported that members in their household wash hands using water from a tap or sink in their dwelling or yard, while a significant number (31 percent) of households relied on a bucket to wash their hands, and only a 7 nominal number (3 percent) of households reported not having a place to wash hands. Using a tap for hand-washing was far more common in urban areas (69 percent) compared to rural areas (46 percent). Of the households that lacked soap, a majority (93 percent) stated they were unable to afford soap with only minimal (2 percent) of households citing a lack of soap availability. A majority of respondents (60 percent) from households with school-aged children (aged 6-14) said that all or some children in their household completed some education at home during the school shutdown. The majority of respondents from households with school-aged children said they usually attended school before the lockdown (88 percent). The most common location for children during lockdown was at home, with most households expressing concern about these arrangements. Apart from the initial school lockdown from March-May in 2020, schools have operated as normal with the exception of brief closures within the Honiara emergency zone during “practice lockdowns”. The most common challenge for children with home learning was lack of access to textbooks at home, followed by availability of worksheets. Only a few households (9 percent) with school aged children reported that the children in their household received some form of remote learning support. Around four fifths (83 percent) of respondents from households with school aged children expressed concern about children, or children’s performance and learning during the lockdown. These findings suggest that there are educational costs to home schooling or, more starkly, risks of learning loss. This concern is in addition to the imposition on adults in the households from children’s needs for supervision and educational support. Analysis of the World Bank HFPS did not demonstrate a widespread perception of a deterioration in community trust and social relations in the second half of 2020. This report focuses on community trust and social areas covering from July to December 2020, with further monitoring planned, as much may have changed in 2021. Overall, a majority (66 percent) of respondents indicated that trust and social relations within the community had remained the same in the six months prior to the survey, with slightly less than one-quarter (22 percent) saying relations had improved and fewer (13 percent) indicating that they believed trust and social relations had deteriorated. Gender differences in views of safety and security related issues followed different patterns in urban and rural areas, with women in rural areas being more likely than women in urban areas to report a deterioration in the safety situation in their community regarding domestic abuse. Of respondents who participated in both Round One and Round Two of the HFPS, significant shares indicated deterioration in security issues such as theft, across both the first half and second half of 2020. These data are subjective and lack a baseline as a point of reference: the results must be interpreted with caution. 8 Executive Summary Given the results of the analysis, there are several areas worthy of policy consideration, in relation to economic recovery, protection of the vulnerable, public services, and public information. Employment growth will need to continue and become more widespread geographically and across the welfare distribution so that rural households and those in the bottom 40 percent are included. The Government of Solomon Islands could potentially promote employment growth through business support and labor-intensive public investment projects. Protecting the socially vulnerable through expanded outreach and services to address domestic violence also seems warranted. In terms of public health, policy makers could seek ways of expanding the supply of clean drinking water and expanding the distribution of COVID-19 vaccines. The COVID-19 vaccine rollout needs to be complemented by better community communication of targeted COVID-19 vaccine information (regarding safety and side-effects) to reduce vaccine hesitancy. More generally, the targeting of public information could also be a consideration for economic policies such as the ESP, to make sure the potential beneficiaries of any policy are adequately aware of available options. 9 1. The COVID-19 situation in Solomon Islands As of April 2021, Solomon Islands remained largely COVID-19 free, with 20 confirmed cases since the start of the pandemic2. All 20 of the confirmed cases were imported during Solomon Islands Government’s (SIG) repatriation program and successfully contained through the country’s quarantine and isolation system3. Essential cargo flights operates on a regular (usually weekly) basis while international trade continues under strict containment measures at all ports of entry. Yet, the threat of importation of the virus remains high, with sudden outbreaks and new strains of the virus appearing in other countries, while illegal border crossings, particularly over the Western Border with Papua New Guinea, are a concern. In the absence of high take up of COVID-19 vaccines, an outbreak of COVID-19 in Solomon Islands would be devastating for an already challenged health system and economy, which has been affected by the global restrictions on international travel, affecting, among others, natural resource exports, the tourism industry, and international remittances. A State of Public Emergency (SOPE) has been in effect and been extended several times since 25 March 2020, to ensure the SIG COVID-response remains operational as the SIG decides on a forthcoming Public Health Emergency Bill. When the international COVID-19 outbreak began in March 2020, the SIG enacted a series of response measures including restrictions on movement, lockdowns and the closure of schools, most markets, and entertainment venues. Most schools and business entities (except for nightclubs in Honiara at the time) reopened again in May 2020. With no community transmission to date, Solomon Islands has since avoided the need to impose strict lockdowns as seen in other countries. Even as two imported cases of COVID-19 were confirmed on 13 October 2020, businesses, schools sporting events 2 Seventeen of the 20 total confirmed cases were recorded between October 3 and November 25, 2020. 3 Repatriation flights running from May 26, 2020 to January 10, 2021 saw over 1,700 residents and foreign nationals entering the country since international travel restrictions were introduced. 10 1. The COVID-19 situation in Solomon Islands and churches remained open, subject to the full observance of COVID-19 safety measures. The one notable exception being for early childhood classes in Honiara, which were suspended until 19 October 2020. Informal markets that operated near the SIG quarantine facilities were also forced to close by the SIG. To prepare for the potential need for lockdowns in the event of a future outbreak, SIG has held “practice lockdowns” in Honiara which only last for a short time (1-3 days). Solomon Islands has continued to receive significant support, as of April 2021, from donors such as Australia, New Zealand, Japan, and China, as well as development partners including the WHO, UNICEF, ADB, IMF, and the World Bank. In addition to ongoing technical assistance and the supply of COVID-19 vaccines (discussed below), key support packages delivered to Solomon Islands since the World Bank High Frequency Phone Survey Round One report (Round One) include the following: USD 0.10 million (Tranche 2) and USD 0.14 million (Tranche 3) in debt service relief from the IMF; USD 5 million in World Bank financing through the COVID-19 Emergency Response Project; USD 0.15m from China in support of the vaccination rollout, operations, and logistics; and the delivery of cash grants and resources in all Solomon Islands provinces as part of Australia’s AUD 13 million COVID-19 budget support package, of which AUD 8 million is allocated to health and 5 million to economic stimulus. On 22 December 2020, Australia and the WHO handed over additional GeneXert calibration kits and cartridges to support the expansion of COVID-19 testing4. Solomon Islands vaccine rollout began on 24 March 2021. The first shipment of 24,000 Oxford-AstraZeneca vaccines arrived on 19 March 2021 via the COVAX Facility5 with technical support from UNICEF and WHO. The COVAX goal for Solomon Islands is to deliver vaccines for 20% of the population by the end of 2021. Although the rollout was underway by April 2021, vaccine hesitancy and wastage were a concern – due in part to misinformation around COVID-19 vaccine safety and effectiveness, high risks of expiration, breakage, and heat exposure (particularly in rural areas that lack proper refrigeration facilities). Vaccine shipments arriving through April 2021 also included consignments from China on 12 April 2021 (50,000 Sinopharm vaccines) and Australia (the first 13,000 of 60,000 committed Australia-AstraZeneca vaccines). At the time the UNICEF SIAS data collection concluded, the vaccination rate was still below 1 percent, therefore the findings present in this report reflect the views and experiences of an almost completely unvaccinated population. 4 Supporting the expansion of COVID-19 testing to Kilufi hospital, Malaita province, Taro hospital in Choiseul, and the National Public Health Laboratory at the Solomon Islands National University in Honiara. 5 COVID-19 Vaccines Global Access (COVAX) is a global initiative to provide equitable access to COVID-19 Vaccines, directed by the Gavi Alliance and is funded principally by donor countries. The World Bank Group is working with COVAX, WHO, and UNICEF to support countries in making payments to COVAX, including the purchase of additional doses beyond the 20% provided by COVAX as countries aim for higher levels of coverage. 11 Figure 1. Population in Solomon Islands that has received at least one dose of the COVID-19 vaccine 8% 7% 6% 5% 4% 3% 2% 1% 0% 3-23-2021 4-23-2021 5-23-2021 6-23-2021 7-23-2021 Source: COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University; Oxford COVID-19 Government Response Tracker, Blavatnik School of Government; via Our World in Data While Solomon Islands has not suffered a health crises from the COVID-19 pandemic, border disruptions, weak external demand, and precautionary health measures resulted in GDP contracting by an estimated 4.3 percent in 20206. During this period there was a significant drop in primary sector production, including in fishing and logging. Furthermore, the severe drop of international visitor arrivals impacted the tourism sector: visitor arrivals dropped from 8,891 in the third quarter of 2019 to 96 in the third quarter of 2020. Large, donor-funded infrastructure projects were delayed due to travel restrictions, therefore dampening construction activity. Solomon Islands Government (SIG) foresaw substantial economic consequences and responded with an ESP. Launched in May 2020 the package was designed to support households and businesses during COVID-19 at a value equivalent to approximately SBD309 million or 2.6% of GDP. More details of the ESP are provided in Section 5. World Bank East Asia and Pacific Macro Poverty Outlook, Annual Meetings 2021. 6 12 I. Background 2. Data collection This section describes the survey methods and data underlying this report. The data was collected through two surveys: Round Two of the HFPS conducted by the World Bank and the SIAS conducted by UNICEF. As both surveys used overlapping samples, they were analyzed together, and the results are presented in this joint report. For the World Bank Round Two survey, as with the Round One data collection, all survey instruments and procedures were designed in accordance with the best practices laid out by the World Bank’s COVID-19 Methodology and Measurement Task Force7. In addition to the information below, further details are provided in the technical appendix and in the Round One report8. 2.1. Description of Survey Objectives & Instrument The objective of Round Two of the World Bank HFPS (conducted in Dec 2020-Jan 2021), was to measure the continued socioeconomic impacts of COVID-19 in Solomon Islands, including on livelihoods, food security, and public safety and security. Due to the difficulty of re-contacting the Round One sample, the majority of respondents in Round Two were new respondents (details provided in the following section). The length of the survey was limited to a target of 15 minutes and the survey instrument consisted of the following modules: Basic Information, Employment and Income Loss, Coping Strategies, Health, Public Trust and Security, Assets, and Wellbeing, with the Health and Public Trust and Security modules being asked to a random 50 percent subset of survey respondents. The questions on employment and income were asked of the respondent. In addition, if the respondent was not themselves the head of households, the respondent was asked to supply information for the head of household. The recall period for current employment was in the week previous to the survey. In addition, retrospective questions were asked for new 7 The five volume best practice methodological guidelines are available at https://documents.worldbank.org/en/publication/docu- ments-reports/documentlist?repnb=148213. 8 The Round One report is available through the World Bank’s Open Knowledge Repository at http://hdl.handle.net/10986/34908. 13 respondents about the baseline (“the start of this year 2020”) as well about the situation at the time of Round One in July 2020 (“around the time of the Independence Day holiday”). The information from the new respondent was then pooled with the returning respondents to have three consistent points in 2020. Three subsequent rounds are planned, with the next completed by August 2021. The implementation calendar may be revised to respond to changing conditions on the ground. A companion survey sponsored by UNICEF was fielded (April 2021) attempting to recontact all of the Round Two households. The objective of the companion survey was to complement the World Bank’s HFPS measuring the socioeconomic impacts of COVID-19 in Solomon Islands with additional information on children related indicators and information. The length of the survey was limited to 15 minutes and the survey instrument consisted of the following modules: Access to Health Care; Employment and Income; Family Arrangements: Child Care, Parenting, Playtime and Mobility; Impact on Education; Mental Health; Water and Sanitation; and Handwashing. This survey can be linked to the World Bank survey through household-level identifiers and therefore the two data sets can be analyzed together. 2.2. Methodology and Data Collection Telephone interviews were conducted through a Solomon Islands call center set up by Tebbutt Research. The dates of World Bank HFPS Round One implementation were June 20 through July 4, 2020. The dates of implementation for the World Bank HFPS Round Two were between December 10 and December 23, 2020, and January 13. The dates of implementation of the UNICEF SIAS were between April 2nd, 2021 and April 21, 2021. As the objective of the survey was to measure changes as the pandemic progresses, Round Two data collection sought to re-contact all 2,665 households contacted in Round One. The protocols for re-contact were a maximum of 3 attempts per caller shift, spaced between 1.5 and 2.5 hours apart depending on whether the phone was busy or there was no answer, and 15 attempts in total. Of the Round One households, 1,048 were successfully re-contacted. To reach the target sample size of at least 2500 households, 1,833 replacement households were added to the World Bank survey. The majority of these were replaced through Random Digit Dialing, but the project did attempt to leverage contact information from ward-level focal points for the Rural Development Project (RDP) in provinces underrepresented in Round One. Of the 145 RDP contacts provided to the call center, 41 were reached, who in turn provided 379 numbers which were attempted as part of regular call schedule. Overall, the sample size achieved for the second round of the HFPS was 2,882 households. As in Round One, the employment questions were asked for both the respondent and the household head by proxy (if different from the respondent), yielding a total sample size for the individual-level employment analysis of 4,279. 14 2. Data Collections Re-contact was attempted with all households from the World Bank Round Two HFPS sample, by phone, for follow up interviews for the UNICEF SIAS. Up to 5 re-contact call attempts were made per house, resulting in 1530 households being interviewed successfully including households without children. Of these households, a total of 1197 had at least one child (aged 0 to 14 years of age). While the goal was to recontact at least 1500 households with at least one child in the household, this was not possible due to lower than hoped for response rate. Given the time elapsed between the Round Two HFPS and the UNICEF SIAS, the response rate may have suffered because of some households changing phone numbers. Due to the high turnover in respondents between Round One and Round Two of the World Bank HFPS and the need for replacement with new households in Round Two, the statistics reported for employment and income differ from statistics reported in the Round One report. The returning respondents in Round Two were initially asked in Round One about their employment status in January 2020 and about their employment status in June 2020. These respondents were then asked again in Round Two about their employment status in June 2020, and also about their employment in December 2020. For the new respondents in Round Two, retrospective questions about employment in both January 2020 and June 2020 were asked, in addition to asking about their current employment status (December 2020). The weighted percentages for June are very close to the Round One estimates. However, the pre-pandemic (i.e. January 2020) estimates are significantly different. More information is provided in Section 3 of this report but it is unfortunately not possible to determine which set of employment statistics is more accurate. The improved coverage of the poorest deciles in Round Two could be considered to be more likely to reflect the true situation of employment. However, new respondents in Round Two were asked to think back to their employment situation one year prior to the survey, which is a substantial recall period that may be associated with significant recall error. 2.3. Re-weighting Despite geographic quota targets, re-weighting was necessary to compensate for areas where targets were not reached. As in Round One, Honiara was over-represented in the World Bank HFPS, constituting 32.8 percent of the survey sample. All other provinces were under-represented, with the largest differences being for Makira-Ulawa, which represented 3.9 percent of the survey sample compared to 7.2 percent of the population in the census, and Guadalcanal, which represented 14.3 percent of the survey sample compared to 21.4 percent of the population in the census. Compensating factors for these differences were developed and included in the re-weighting calculations. See Appendix 1.2 and Figure 43 for further information. 15 Re-weighting was also necessary in both surveys to adjust for demographic and wealth characteristics of households. Weights are required for unbiased estimation. In addition to the geographic oversampling above, because the survey was administered by mobile phones, the respondents were a representative sample of mobile phone holders, not the population overall, and non-random non-response can exacerbate these differences. The Round One report found respondents were more likely to be male, urban, wealthier, and more highly educated. To make inferences at the level of the population instead of mobile phone holders, it was necessary to reweight the survey data. Details on this process are provided in the technical appendix showing the pre weighting and post-weighting distributions of the main demographic characteristics. See Figure 44 and Figure 45. In the appendix for further information on the pre-and post-weighting distributions and also the Round One report. Given that there was an underrepresentation of households from lower deciles and some provinces in Round One, it was expected that this would also be the case, for Round Two. However, the inclusion of rural households found through the Rural Development Project (RDP) resulted in a sample that was more balanced than in Round One, albeit one that is still not representative of the wealth deciles or provinces, in the absence of appropriate weights. 2.4. Analytical approach Both descriptive statistics and econometric analysis were used to examine relationships in the data. Unless otherwise specified in the text, the variables included in the regression at the level of the individual were sex, age, status as head of household, location (Honiara, other urban, and rural), wellbeing status/wealth quintile (bottom 40 percent, third quintile, fourth quintile, fifth quintile/top 20 percent), pre-COVID-19 sector of employment (agriculture, industry, services), and education (no formal education, some or completed primary, some or completed secondary, tertiary, vocational, and other). At the household level, the included variables were location (province and urban or rural location) and wealth quintile. For the reporting of descriptive statistics, many of the charts throughout the report provide disaggregation by sex of head of household, (or the sex of respondent), whether or not the households have at least one child under 15, location (Honiara, other urban, and rural), and wellbeing status/wealth quintile (bottom 40 percent, third quintile, fourth quintile, fifth quintile/top 20 percent)9. When discussing statistical point estimates, confidence intervals are reported in the text in brackets e.g. (CI: 17.4, 29.8). The levels of disaggregation reported are not always indicated in the chart heading, but groups are clearly labeled within charts. 9 16 3. Employment and Income This section reports analysis of changes in employment and income10. It begins with estimates of employment for three points in time across 2020- January, June and December, and how this varies across groups and industries and the formal and informal sectors. This section also reports analysis of income changes, analysis of agricultural activities, and analysis of remittances. It concludes with analysis of households’ access to financial services and the economic outlook of respondents. 3.1. Employment Despite some recovery between June and December 2020, employment had not recovered to pre-pandemic levels, following a fall in employment from January to June 2020. As discussed in Section 2.1, there are differences in the employment statistics from Round One and Round Two but in the absence of some other representative data set for reference, it cannot be determined which set of statistics is more accurate. However, both new and returning households showed the same pattern in overall employment: a decline between January and June, followed by a partial recovery between June and December. In Round One, 43.6 percent reported working in January (CI: 34.7, 52.8) and 40.5 percent (CI: 32.0, 49.6) reported working in June. In Round Two, 58.1 percent reported working in January (CI: 52.5, 63.6), 45.0 percent reported working in June11 (CI: 39.6, 50.4), and 47.1 reported working in December (CI: 41.6, 52.5)12. While there is little difference between June and Dec in the national employment rate, Figure 2 shows considerable “churning” in an out of employment. Many who were working in June were out of work in Dec; and many not working in June had found work in Dec. This means that the headline numbers mask some level of volatility. Nearly one-fifth of Round Two 10 In section 3, employment and income analysis uses individual level data that is based on data for both respondents and the head of household. For analysis of non-farm business, agriculture, remittances, access to financial services, and financial anxiety, this section uses household level data. 11 The employment question for new respondents asked them to think back to “around the time of Independence Day Holiday” (July 7th). The data is therefore indicative of employment at the end of June and the start of July, rather than the month of June (or the month of July). The data is consistent with Round One data which was collected from June 20 through July 4, 2020. 12 The underestimation of those working in January in Round One is consistent with the findings from a multilevel regression and stratification (MRP) analysis. However, even the modeled estimates are below the Round Two estimates of baseline employment. Unless otherwise noted, employment statistics refer to data from the Round Two survey questions. 17 respondents lost employment between January 2020 and December 2020. This response may overstate job losses due to temporary holiday closures or vacation. A substantial share of Round Two respondents at 45.7 percent, (CI: 40.1, 51.3) transitioned into or out of work at least once over the course of 2020 with 27.3 percent (CI: 22.5, 32.0) reporting not working in either. Of those that stopped working between June and December, the main reason cited was vacation or holiday closures, indicating that the partial recovery may have been more pronounced than are reflected in these numbers. Figure 2 below shows the movement between work status across January, June and December 2020, based on Round Two data. Figure 2. Change in work status by round (January 2020- July 2020- Dec 2020) January June December working working not working not working Source: Round two of the World Bank high frequency mobile phone survey Note: The two work status categories are mutually exclusive and account for all respondents in Round Two as for each time period a respondent could have reported either working or not working. Each grey-shaded line traces the group of respondents from one work status in a time period, to either a different or the same work status in the next time period. The thickness of the line represents the proportion of respondents moving from one work status to the other, or staying in the same work status. The magnitude of the employment recovery differed across groups (see Figure 3). Urban households demonstrated the strongest employment recovery, while households in the bottom 40 percent fared poorly. For those living in Honiara, and other urban areas, around a quarter of the initial decrease seen between January and June had recovered by December. In rural areas, the losses from January to June were similar in magnitude, but less than 5 percent of these jobs had recovered by December. For those in the bottom 40 percent, there had been no significant recovery from June to December, with a 25.5 percent net loss over 2020. This compares to a net loss from Jan-Dec 2020 of 10.4 percent in the third quintile, 8.4 percent in the fourth quintile, and 19.3 percent in the top 20 percent. Only the fourth quintile experienced significant employment recovery from 46.1 percent (CI: 38.5, 53.8) in June to 55.8 percent (CI: 47.8, 63.5) in December. For men, there was a fall in employment from 62.3 percent (CI: 56.1, 68.2) in January 2020 to 49.9 percent (CI: 43.8, 56.1) in June and no significant recovery by December. For women there was a 18 3. Employment and Income fall in employment from 54.1 percent (CI: 45.1, 62.9) in January 2020 to 40.4 percent (CI: 32.2, 49.2) in June and slight but insignificant recovery by December to 43.9 percent (CI: 35.4, 52.8). Employment rates for households with children13 were not significantly different from employment rates for households without children in December but followed different trajectories across 202014. Employment in the service sector faced a greater net decline than employment in agriculture, from January 2020 to December 2020 (see Figure 3). The service sector employed 35.3 percent (CI: 30.3, 40.6) of adults in January, fell to 24.9 percent (CI: 20.7, 29.5) in June and recovered to 27.6 percent (CI: 23.3, 32.4) in December. In contrast, the agriculture sector employed 18.5 percent (CI: 14.7, 23.0) of adults in January, fell to 16.8 percent (CI: 12.9, 21.6) in June and then was almost the same in December at 15.8 percent (CI: 12.0, 20.5). Since agriculture is the main sector of employment for rural residents, women, and those in the bottom 40 percent, understanding the impact of COVID-19 on agricultural activities is critical to understanding its impact on poor and vulnerable groups15. Agriculture is further explored in section 3.4. The small industrial sector did not show a statistically significant increase. Figure 3. Percentage of adults in employment in January 2020, June 2020 and December 2020 men women Honiara Other urban Rural No children At least one child 70% 70% 70% 60% 60% 60% 50% 50% 50% 40% 40% 40% 30% 30% 30% January June December January June December January June December bottom 40% 3rd quintile 4th quintile top 20% Agriculture Industry Services 70% 40% 60% 30% 50% 20% 40% 10% 30% 0% January June December January June December Source: Round Two of the World Bank high frequency mobile phone Note: Data for December was in reference to the past week while recall data was used for January and June for employment at “the start of the year” and “around the time of Independence Day”, respectively. 13 At least one child below 15 years of age. 14 There was sharper drop in employment for households with children in the first half of the year that saw the rates diverge in June. 15 See Round One report for further details: http://hdl.handle.net/10986/34908. 19 Jobs in the formal sector in January 2020 showed a stronger recovery than jobs in the informal sector. Based on the Round Two recall data of those employed in January 2020, 23.1 percent (CI: 18.7, 28.2) were in the formal sector and 76.9 percent (CI: 71.9, 81.3) were in the informal sector. Figure 4 is based on employment levels in January 2020 as a baseline, reporting the percentage of those employed in June and December as a share of those employed in the January 2020 baseline by employment type. While the formal sector declined in June 2020, it recovered to 94.6 percent baseline levels by December 2020. Informal sector employment in December 2020 was only 76.2 percent of what it was at baseline. Both sectors had similar levels of downturn in June, but there has been a stronger recovery for those in the formal sector. Part-time jobs at baseline (accounting for 29.5 percent of all employed in January, CI: 23.5, 36.4) saw a reduction of almost 50 percent between January and June. Relative to full-time employment this fall was much greater. However, part-time employment experienced a significant recovery between June and December while full time employment has not significantly recovered. Both part-time employment and full-time employment remain well below pre-pandemic levels. This data suggests that businesses initially coped with the economic slowdown by downsizing their part-time staff more so than their full-time staff. Figure 4. Changes in employment by sector and job type, relative to a January 2020 baseline formal informal full time part time 100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% January June December January June December Source: Round two of the World Bank high frequency mobile phone survey. While there were significant flows into and out of employment across 2020, switching between industrial sectors remained uncommon. Figure 516, shows the change in activity and sector across Rounds One and Two. Overall, 87.4 percent (CI: 82.8, 90.9) of those employed in January, June and December, were engaged in the same activity and 92.7 percent (CI: 89.1, 95.2) remained in the same sector. Between January and June, 5.8 percent (CI: 3.8, 8.8) of respondents reported switching activities compared with 14.4 percent (CI: 10.4, 19.6) between June and December. The sectors and activities were defined as agriculture and logging activities in the agricultural sector: mining, manufacturing, electricity, 16 water, gas, waste, and construction activities in the industry sector. Further identified were, professional, scientific, technical, transportation, retail, wholesale, trading, finance, insurance, real estate, personal services, education, health, public administration, tourism, restaurants, hospitality, handicrafts, and cultural industries, undefined self-employment, security and defense, community works and service, and other activities in the service sector. 20 3. Employment and Income Between January and June, 2.5 percent (CI: 1.3, 4.9) switched sectors, switching activities and 6.6 percent (CI: 4.3, 10.0) switching sectors between June and December. In the absence of detailed employment data, it is not possible to determine if the changes were due to individuals seeking alternative employment because of COVID-19 related disruptions or because of natural cycles in employment, such as those around the agricultural calendar. However, as the majority of those that switched back were in the service sector, it is more likely to have been the former rather than the latter. Figure 5. Change in sector of employment (by round) January June December not working not working agriculture agriculture service service industry industry Source: Round Two of the World Bank High Frequency Mobile Phone Survey Note: The four “sectors”, which include the category of not working, are mutually exclusive and account for all respondents in Round Two. So for each time period, a respondent could have reported being in any sector category. Each grey-shaded line traces the group of respondents from one sector of employment in a time period, to either a different or the same sector in the next time period. The thickness of the line represents the proportion of respondents moving between the two sectors connected by the line (or staying in the same sector if the line connects the same two sectors). For example, the thickest line is for the group that was not working in June, and also not working in December. This line is the thick straight line in the top right area of the chart. 3.2. Income The income shock has been more widespread than the loss in employment, and remains below pre-pandemic levels with little recovery post-June 2020. Figure 6 reports changes in income for data from Round One and data from Round two. Respondents were asked about income earned from any activity including business, employment and agriculture. In Round Two, 11.8 percent (CI: 7.9, 17.2) of respondents reported increased income in the previous week, 43.1 percent (CI: 35.8, 50.8) report 21 the same level of income, 37.3 percent (CI: 30.0, 45.2) reported lower income, and 7.8 percent (CI: 3.3, 17.4) reported no income. These percentages were not significantly different from Round One. For both those living in Honiara and other urban areas, the percentage reporting lower than usual income was lower in Round Two than Round One. In contrast, rural households were more likely to report lower than usual income in Round Two (38.8 percent, CI: 29.5, 49.0) than in Round One (29.5 precent CI: 19.0, 42.6). For those in the top quintile, the percentage reporting higher than usual income was greater in Round Two than Round One. Most respondents reported being able to work as normal in Round Two17. Figure 6. Changes in usual income between Rounds One and Two, by location Increased Remained the Same Decreased No Payment 60% 40% 20% 0% 20% 40% 60% 80% r1 r2 r1 r2 r1 r2 r1 r2 Honiara Other Urban Rural National Location Increased Remained the Same Decreased No Payment 60% 40% 20% 0% 20% 40% 60% 80% r1 r2 r1 r2 r1 r2 3rd Quintile 4th Quintile Top 20% Note: The Round One sample for those in the bottom 40 is very small yielding estimates of low precision. Therefore, estimates are not reported for the bottom 40. Source: Rounds One and Two of the World Bank High Frequency Mobile Phone Survey The percentage of those who could work normally in Round Two was high at 82.7 percent, but is lower than in Round One (93.1 17 percent). This finding may be related to the timing of Round Two data collection during the holiday period. Income data is at the individual level, not household level. 22 3. Employment and Income 3.3. Non-Farm Business 18 Compared to Round One, there has been a marked deterioration in income from non-farm enterprises. Figure 7 compares the change in household income from non-farm enterprise in the month prior to data collection between Rounds One and Two by location. Nationally, the share of those households with non-farm enterprises indicating income was the same as usual fell significantly from 56.2 percent (CI: 39.1, 71.9) in Round One to 33.4 percent (CI: 17.1, 55.0) in Round Two, with those households reporting less than usual income rising from 31.8 percent (CI: 19.9, 46.6) to 50.8 percent (CI: 31.1, 70.2) between Rounds One and Two. Little change was reported by respondents reporting no income with a slight increase from 3.3 percent in Round One to 5.0 percent in Round Two. Those reporting higher incomes reported from 8.7 percent (CI: 3.1, 22.3) in Round One to 10.8 percent (CI: 6.1, 18.2) in Round Two. Rural households were more likely to report lower income than usual in the past week in Round Two (47.9 percent CI: 17.5, 80.0), compared to 28.2 percent (CI: 15.9, 45.0) in Round One. In contrast, households in Honiara were not more likely to report less than usual income in the past week in Round Two compared to Round One. Overall, changes in non-farm income are broadly consistent with changes in income more generally (as presented in section 3.2). Figure 7. Change in income from non-farm enterprise in last month compared to start of 2020 (Round One and Round Two), by location and well-being status Higher than usual Same as usual Less than usual No income 60% 40% 20% 0% 20% 40% 60% 80% r1 r2 r1 r2 r1 r2 r1 r2 Honiara Other Urban Rural National Region Source: Rounds one and two of the World Bank high frequency mobile phone survey In Round One, 27.1 percent (CI: 17.8, 36.4) of households reported operating a non-farm business in 2020. In Round Two, 24.5 percent 18 (CI: 11.8, 43.9) reported having a non-farm enterprise. 23 3.4. Agriculture Nationally, 40.5 percent (CI: 28.9, 52.1) of households participated in agricultural activities in the month prior to the survey19, with 87.5 percent (CI: 78.3, 96.8) being able to perform these activities normally. In terms of the composition of agricultural activities, the production of food crops primarily for home consumption was the most common main agricultural activity for households, cited by 60.9 percent (CI: 42.2, 76.9) of agricultural households, followed by the production of crops for sale in local markets, which was cited by 19.1 percent (CI: 8.1, 38.8) of agricultural households. Other main activities included fishing for sale (9.0 percent, CI: 2.0, 32.9), raising livestock for sale (7.7 percent, CI: 4.7, 12.2), production of cash crops for sale (1.6 percent, CI: 0.6, 4.3), and production of livestock for home consumption (1.1 percent, CI: 0.4, 2.7), with a negligible share citing fishing for home consumption. More than four-fifths of agricultural households consumed at least part of their production (83.4 percent; CI: 75.0, 89.3) while less than half (CI: 22.2, 56.6) sold their products in local markets. Households in the bottom 40 percent of the wealth distribution were the most likely to produce crops for sale, 41.8 percent (CI: 15.3, 74.1), compared to 34.8 percent (CI: 19.1, 54.6) of households in the third quintile, 38.4 percent (CI: 25.0, 53.8) of households in the fourth quintile, and 28.1 percent (CI: 19.6, 38.5) of households in the top 20 percent. Across the regions, 23.6 percent (CI: 12.8, 39.3) of households in the Honiara region, 37.4 percent (CI: 25.4, 51.1) of households in other urban areas, and 41.8 percent (CI: 20.6, 66.5) of households in the rural areas produce crops for sale in the local market. Econometric results confirm that those who are in the third quintile were about 27 percent more likely to do so than the top 20 percent, but no other characteristics were significant. See Table 3 in the appendix for further information. Due to the sample being mostly urban, and the difficulty of obtaining rural respondents, the survey cannot provide precise estimates 19 of the levels of agricultural production and the percentage sold. 24 3. Employment and Income The purpose of agricultural production (subsistence versus income) varies by the sex of the head of household. Figure 8 summarizes descriptive statistics of the main agricultural activities across household characteristics. Female-headed households were more likely to cite production of food crops for home consumption as their main agricultural activity, in contrast to male-headed households who were more likely to cite production of cash crops for local market as indicated by econometric analysis to hold other factors constant. See Table 6 in the appendix for further results. Male-headed households were more likely to report fishing for income, while fishing was almost non-existent for female-headed households. While both male-headed households and female-headed household use agriculture for subsistence, this is more common for female-headed households, being the main activity of more than four-fifths of all female-headed agricultural households. The purpose of agricultural production also varies between households with and without children. Households without children are more likely to cite the production of fish for sale (32.6 percent, CI: 5.8, 78.7). Figure 8. Main agricultural activity, by sex, family structure, location, and well-being status Food crops for home consumption Crops for sale in local market Cash crops for sale or export Livestock for home consumption livestock for sale Fishing for home consumption Fishing for sale 100% 80% 60% 40% 20% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child urban 40% Quintile Quintile National Sex of HH head Family structure Location Well-being status Source: Rounds one and two of the World Bank high frequency mobile phone survey Almost half of households expect lower or no income from agricultural activities compared to the last agricultural season. Overall, only 54.2 percent (CI: 34.2, 73.0) of households expect to see higher or the same income from their agricultural production in the current growing season compared to the last agricultural season, as shown below in Figure 9. Female-headed households were more optimistic about agriculture earnings, with 24.8 percent (CI: 9.6, 50.7) expecting lower or no income compared to 56.0 percent (CI: 35.5, 74.6) of male-headed households. Households in the third and fourth quintile of the distribution were more likely to expect lower or no earnings from agricultural production, 49.2 (CI: 34.6, 64.0) and 56.0 percent (CI: 32.4, 77.2), respectively, compared to 43.5 percent (CI: 33.6, 53.9) of households in the top 20 percent of the wealth distribution and 42.7 percent (CI: 13.9, 77.4) in the bottom 40 percent. Limited variation was detected between urban and rural areas. 25 Figure 9. Expected Agricultural Earnings in the Current Season, by sex, family structure, location, and well-being status Higher Than Usual Same As Usual Less Than Usual No Income 60% 40% 20% 0% 20% 40% 60% 80% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH Head Family structure Location Well-being status Source: Round two of the World Bank high frequency mobile phone survey While one-third of households cited higher demand for agricultural products, a larger proportion cited lower demand. Overall, 33.4 percent (CI: 11.5, 65.8) of households indicated higher demand for their agricultural products than one year prior to the survey. In contrast, 41.1 percent (CI: 15.9, 72.0) of households report lower demand for agricultural products than a year ago. Around a quarter of households cited no change in demand (24.0 percent, CI: 11.7, 42.9). Households citing higher demand may be related to increased prices and decreased availability of some imported foods due to COVID-19. However, given that different households grow different crops, the impact of food price changes would not be even across households. Raising livestock for sale was more common than raising livestock for home production, but was mainly considered a secondary activity to crop production. Nationally, 27.3 percent (CI: 11.8, 42.7) of households engaged in livestock production. Of the households engaged in these activities, 15.2 percent (CI: 6.3, 32.3) raised livestock only for home consumption, 27.5 percent (CI: 13.1, 48.9) only for sale, and 57.3 percent (CI: 30.4, 80.5) for both. In livestock producing households, 74.9 percent (CI: 54.9, 88.0) also engaged in crop production, and the majority considered crop production to be their primary activity. Livestock activities were more common in the rural areas (29.7 percent, CI: 13.1, 54.2), compared with 22.8 percent (CI: 11.9, 39.0) in the Honiara region and 18.5 percent (CI: 10.7, 30.1) in the other urban areas. Few agricultural households engaged in fishing20, either for home consumption or sale. Only 6.1 percent (CI: 0.7, 11.4) of agricultural households reported engaging in fishing activities, and those who did participate were mainly in rural areas. Of those that did engage in fishing, 15.3 percent (CI: 5.1, 37.7) fished only for home consumption, Fishing was not specifically defined in the survey questionnaire and therefore could cover a broad range of activities from shallow 20 water still/bait fishing to fishing using a boat or, canoe, etc. 26 3. Employment and Income 7.9 percent (CI: 2.3, 24.2) only for sale, and 76.8 percent (CI: 48.3, 92.2) for both. Of those engaged in sales, 44.0 percent (CI: 8.2, 87.3) have experienced an increase in demand since this time last year, 9.9 percent (CI: 2.7, 30.9) have experienced no change, and 45.6 percent have experienced a decline. Given the small sample sizes, no further disaggregation was possible. 3.5 Remittances Most remittances from international sources have declined or stopped in the past month relative to what households usually receive, while domestic remittances were more likely to have stayed at the same level (see Figure 10 and Figure 11). Overall, 15.4 percent (CI: 9.4, 24.3) households received domestic remittances and 3.1 percent (CI: 2.2, 4.2) of households received international remittances. 18.0 percent (CI: 11.6, 26.8) of households reported either receiving domestic remittances, international remittances or both. Households in urban areas outside of Honiara were the most likely to receive domestic remittances (30.0 percent, CI: 22.5, 38.1), while those in Honiara were the most likely to receive international remittances (9.4 percent, CI: 6.9, 12.7). Households were asked if the level of remittances they received over the past month were different to what they “usually” receive. Overall, around half of households that receive remittances report the level of remittances has stayed the same in the past month as usual (52.0 percent, CI: 31.7, 71.7) while one third (33.6 percent, CI: 19.5, 51.4) have seen a decrease or had their remittance flows stop entirely. Over half of households with international remittances (53.5 percent, CI: 42.4, 64.3) reported remittances to have reduced compared to what they usually receive. This is higher than the percentage of households receiving domestics remittances that report those remittances to have reduced (30.1 percent, CI: 15.6, 50.2). For households with children, only 34.7 percent (CI: 25.3, 45.5) reported the same level of remittances as usual and 48.1 percent (CI: 35.0, 61.4) received lower remittances than usual. Figure 10. Change in domestic Figure 11. Change in international remittances compared to remittances compared to usual levels usual levels Increase Increase 15% 13% Decrease or stop 30% Stay the Decrease or same stop 34% Stay the 53% same 55% Source: Round two of the World Bank High Frequency Mobile Phone Survey 27 3.6. Access to Financial Services A minority of respondents, at 23.0 percent (CI: 17.4, 29.8), expressed the need to go to a bank, ATM, or post office to make a cash withdrawal in the last month before the Round Two survey. Figure 12 shows the distribution of respondents who needed to go to a bank, ATM, or post office to make a withdrawal, by their sex, location, and well-being status. Respondents located in rural areas (at 10.4 percent, CI: 6.0, 17.4) were less likely to express the need to make a withdrawal. Respondents’ well-being status was also associated with the need to make a withdrawal, with those in the top 20 percent being more likely to express this need at 69.7 percent (CI: 65.0, 74.0) than the rest of the well-being distribution. Econometric analysis in Table 9 confirms that respondents located in rural areas were significantly less likely than respondents in other areas to need to go to a bank, ATM, or post office. Of those respondents who expressed the need to go to a bank, ATM, or post office to make a withdrawal, almost all of them successfully gained access, at 97.2 percent (CI: 94.1, 98.7). Figure 12. Share of households expressed the need to go to a bank, ATM, or post office to make a withdrawal, by sex of respondent, location, and well-being status 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex Location Well-being status Source: Round two of the World Bank high frequency mobile phone survey A minority of respondents expressed the need to go to a money agent in the last month prior to the Round Two HFPS survey, (12.2 percent, CI: 8.2, 17.7). A majority of these respondents also needed to go to a bank, ATM, or post office, at 76.0 percent (CI: 57.4, 88.1). The proportion of respondents who needed to go to a money agent across sex, location, and well-being status is shown in Figure 13. Similar to the need to access a bank, ATM, or post office, respondents who lived in the rural areas were less likely to need to go to a money agent, at 5.7 percent (CI: 2.4, 13.0). The need also tended to increase with respondents’ well-being status, with 29.5 percent (CI: 25.6, 28 3. Employment and Income 33.7) of those in the top 20 percent expressing the need to go to a money agent while only 4.7 percent (CI: 1.2, 17.5) of respondents in the bottom 40 percent expressed the same need. No significant difference was found between households with and without children. Figure 13. Share of households expressed the need to go to a money agent, by sex of respondent, location, and well-being status 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex Location Well-being status Source: Round two of the World Bank high frequency mobile phone survey Of the respondents who needed to go to a money agent (12.2 percent), a majority (81.0 percent; CI: 48.0, 95.2) were able to get access within the last month before the Round Two survey. The most common reason for not being able to access a money agent was that the office was closed, with 88.9 percent (CI: 55.1, 98.1) of respondents citing this reason. Figure 14 shows the proportion of respondents who were not able to go to access a money agent based on sex, location, family structure, and well-being Àstatus. Respondents from households in the bottom 40 percent, were significantly less likely than other households to be able to obtain access to a money agent, at 70.9 percent (CI: 23.5, 95.1). Respondents in rural areas, at 47.2 percent (CI: 12.9, 84.3), were less likely to be able to access a money agent. Figure 14. Share of households who were not able to access a money agent based by sex of respondent, location, and well-being status 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex Location Well-being status Source: Round two of the World Bank high frequency mobile phone survey 29 3.7. Financial Anxiety More than three-quarters of respondents felt worried about their household’s finances over the next month (following the survey), with around a third being “very worried”. Overall, 77.4 percent (CI: 66.4, 85.8) of respondents said they were “somewhat” or “very worried” about their finances over the next month. The level of anxiety was consistent across sex of respondent, family structure, location, and well-being status, as can be seen in Figure 15, with no statistically significant differences. Figure 15. Financial anxiety by sex, family structure, location, and well-being status Very worried Somewhat worried Not too worried Not worried at all 100% 80% 60% 40% 20% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex Family structure Location Well-being status Source: Round two of the World Bank high frequency mobile phone survey Most respondents expected the state of the economy would worsen in the next year. Overall, 67.2 percent (CI: 56.0, 76.7) of respondents expressed pessimism about the economic outlook, of which 73.3 percent (CI: 55.4, 85.9) were households with children. Figure 16 shows respondents’ economic outlook across sex, family structure, location, and well-being. Econometric analysis in Table 8 indicates that respondents in households with a female-head of household were significantly more likely to have a negative outlook on the economic situation (81.4 percent, CI: 65.8, 90.8), while the location of households and well-being status were not related to economic outlook. Figure 16. Economic outlook by sex of respondent, family structure, location, and well-being status Much better Somewhat better About the same Somewhat worse Much worse 40% 20% 0% 20% 40% 60% 80% Female Male No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex Family structure Location Well-being Status Source: Round Two of the World Bank High Frequency Mobile Phone Survey 30 4. Household Coping Strategies and Personal Coping Strategies This section describes household level coping strategies over the six months prior to the World Bank HFPS. It then reports the COVID-19 concerns of respondents and individual level coping strategies for adults and children, based on data from the UNICEF SIAS. 4.1. Coping Strategies of Households21 Around 95 percent of households in Solomon Islands reported using at least one coping strategy22 from July to December 2020. Over 60 percent reported their households employed five or more coping strategies23. These coping strategies, however, are difficult to attribute to COVID-19 as there are no pre-pandemic baseline measures. The most common strategies were finding ways to earn extra money (71.6 percent, CI: 61.0, 82.3), spending from personal savings (60.3 percent, CI: 48.5, 72.1) and reducing consumption (either food or non-food) (58.5 percent and 54.9 percent respectively). Table 1 shows the ten most common coping strategies. Table 1. Top 10 most common coping strategies Find ways to earn extra money 71.6% Spend from savings 60.3% Reduce food consumption 58.5% Reduce non-food consumption 54.9% Purchase items on credit 49.6% Receive other assistance from friends or family 47.5% Receive cash or borrow from friends or family 36.2% Delay making re-payments 35.9% Sell livestock 17.8% Reduce the number of children attending school 17.0% Source: Round Two of the World Bank High Frequency Mobile Phone Survey 21 This section considers household level data from the World Bank HFPS. 22 Full list of coping strategies included: Sell assets; Sell livestock; Find ways to earn extra money; Receive cash or borrow from friends or family; Receive other assistance from friends or family; Receive assistance from church or other religious body; Take a loan from a financial institution; Take a loan from an informal moneylender; Purchase items on credit; Delay making repayments; Sell harvest in advance (agricultural households only); Reduce food consumption; Reduce non-food consumption; Spend from savings; Receive assistance from NGO; Receive assistance from a community based organization; Take an advance from an employer; Reduce the number of children attending school (households with school age children only). 23 It should be noted that not all coping strategies are relevant to all households (for example, only households with livestock can sell livestock, and only households with children in school can reduce the number of children in schooling). However, the statistics are reported for all households to highlight what is most commonly occurring across Solomon Islands. 31 Some common household coping strategies imply increased household debt. Of the ten most common coping strategies as shown in Table 1, three lead to increased household debt: purchasing items on credit (49.6 percent, CI: 37.8, 61.4), borrowing from friends or family (36.2 percent, CI: 25.2, 47.2), and delay repayments (35.9 percent, CI: 24.0, 47.8). Of those households who received informal loans from friends and family, a majority (80.6 percent) are expected to repay the loan, and of those, most (53.0 percent) indicated that they would pay interest on the loan. Households appear to have shifted from short to medium-term coping strategies in the second half of 2020 (see Figure 17). Though high turnover of respondents between Rounds One and Two hinders comparability to some degree, identifying broad trends is still valuable for analysis. The frequency of using coping strategies that accrue more debt increased between July and December 2020, (22.9 percent increase for purchasing items on credit, 6.8 percent increase borrowing from friends or family, and 16.0 percent increase for delay repayments). Households were far more likely to report increasing their income (71.7 percent, CI: 59.9, 81.0, compared to 44.9 percent, CI: 31.1, 59.5), delay making re-payments (35.9 percent, CI: 25.0, 48.5, compared to 19.9 percent, CI: 7.5, 43.1) or drawing down on savings (60.3 percent, CI: 48.1, 71.3, compared to 47.7 percent, CI: 32.6, 63.3) in the second round of the survey, which is also indicative of a shift to more medium term consumption smoothing strategies. Short term strategies continue to be commonly used with more than half of households in Round Two reporting that they had reduced food or non-food consumption as a coping strategy in the previous six months, but the rates are similar to Round One. 32 4. Household Coping Strategies and Personal Coping Strategies There was also a three-fold increase in the share of households reporting they had reduced the number of children attending school in their household. Of households with children, 20.7 percent (CI: 13.3, 30.9) chose to reduce the number of children attending school. As this survey was collected partially during school holidays, it may be an artifact of the timing. However, if this statistic reflects a real reduction in the rate of children attending school, this has implications for the long-term development of Solomon Islands. Subsequent rounds will verify the reported changes. Figure 17. Difference between share of households in Round Two and Round One using coping strategy Change between Round One and Two (percentage points) 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% -5.0% -10.0% n em ion ts um n n s s ce O ey um er ily om y it er om an ck ily ol et ng sis ligi utio io io d ed en G nd oy am ho on an am o bo ss at pt pt N vi t cr m pl s le v a iz re stit sc m sa rf f nc ed live nc us ad ay ey n ll or so a g Se so om o n -p ns ns tr on in in fr rg li l fin ds ba Sel om n fr om an ex nd re m co co nd a e o fr st m n i ite nc ie e rn rie g d te od od al rv in en fr ta er a ea at s rm f se ha ak fo fo r om th Sp f n ha o fo n- e m o ce ll as re a st ity fr Se fr rc in Re or no om ay ld du ay e un ow va Pu ce an iv hi el ch w Re ce m r a ad ce n fc D rr om ur a nd du m bo sis Tak sist ro he an a ch co Fi lo Re fr be or s e an a a ak m sh e om lo T fr nu ca a ot fr e e nc ke e e th e iv nc ta Ta iv ce ce ce ta Re du sis Re as Re as e iv e ce iv ce Re Re Source: Round Two of the World Bank High Frequency Mobile Phone Survey Access to safety nets was mainly through informal channels. Almost sixty percent (59.4 percent) utilized at least one informal channel, with non-monetary assistance from friends or family (47.5 percent) being the most frequently cited strategy. Other informal channels included receiving or borrowing cash from family and friends (36.2 percent), receiving assistance from a church or other religious body (14.6 percent), receiving assistance from a community-based organization (10.2 percent) and receiving assistance from an NGO (1.9 percent). Compared across survey rounds, households have relied less on informal safety nets in the second half of 2020 (see Figure 17). This finding is expected given the drawn-out nature of the economic crisis caused by COVID-19 and the limited resources available for communities and families to support one another. 33 4.2. COVID19 Related Concerns and Coping Strategies of Adults and Children24 More than three-quarters (78.6 percent, CI: 68.3, 86.2) of respondents from households with children had COVID-19 related worries, and the most commonly mentioned reasons were being worried about getting sick themselves, children getting sick, or parents getting sick (see Figure 18). Respondents in female-headed households more likely to state they were worried about their parents getting sick (78.1 percent, CI: 54.9, 91.2) than male-headed households (51.3 percent, CI: 34.8, 67.6). Respondents in female-headed households were less likely to state they were worried about children getting sick (70.9 percent, CI: 42.0, 89.2), than male-headed households (79.8 percent, CI: 66.9, 88.6). The majority of respondents who had children in their household were concerned about their parents getting sick (56.0 percent, CI: 40.3, 70.5), and an even greater proportion were concerned about children getting sick (78.3 percent, 66.8, 86.6). Respondents in male-headed households were around twice as likely as respondents in female-headed headed households to state that they were worried about the future of their community or country. The most common emotions respondents from households with children reported experiencing at the time of the UNICEF SIAS were: being happy (51.0 percent, CI: 39.4, 62.5), being anxious or worried (30.3 percent, CI: 21.1, 41.4 percent) and being tired (17.7 percent, CI: 8.6, 33.0) (see in Figure 47 in Appendix). In particular, households in Honiara were more likely to express being worried or anxious (45.7 percent, CI: 32.6, 59.4), compared to those in other urban areas (16.8 percent, CI: 9.1, 29.0) and rural areas (27.4 percent, CI: 16.9, 41.3). Around three-quarters of respondents stated that children were happy (see Table 3 in Appendix). More generally however, it isn’t’ possible to infer an impact of COVID-19 on anything mental health related due to a lack of baseline data25. Figure 18. Covid-19 related worries reported (for households with children only), and by sex of household head, among households with households Other Not knowing how to home-school kids Not knowing how to help kids cope with their stress or worries Not knowing how to talk to kids about COVID Less patience with children More responsibilities Future of community/country Lack of access to services Can not see family/friends Kids education Staying at home Staying in quarantine Losing job Family income Parents getting sick Kids getting sick Getting sick 0% 20% 40% 60% 80% 100% 24 This section uses data from the UNICEF SIAS on the views of respondents from households with children. It consider these respondents’ views on their own emotions and coping strategies, and the coping strategies of children within their household. 25 The Second Round of the UNICEF SIAS will be able to determine if COVID-19 related worries and anxiety, change over time. 34 4. Household Coping Strategies and Personal Coping Strategies Other Not knowing how to home-school kids Not knowing how to help kids cope with their stress or worries Not knowing how to talk to kids about COVID Less patience with children More responsibilities Future of community/country Lack of access to services Can not see family/friends Kids education Staying at home Staying in quarantine Losing job Family income Parents getting sick Kids getting sick Getting sick 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Female-headed household Male-headed household Source: UNICEF mobile phone survey The most common coping strategies employed by households with children were prayer (73.9 percent, CI: 61.0 83.7), followed by staying connected with family/ friends (41.8 percent, CI: 28.7, 56.2), staying positive (31.7 percent, CI: 20.5, 45.4) and reading the news (26.9 percent, CI: 15.6, 42.2). Prayer and reading the news were far more common strategies in male-headed households, whereas staying connected with family and friends and staying positive were far more common coping strategies in female-headed households and those without children (see Figure 19). Households in rural areas were more likely to report relying on prayer, whereas those in urban areas were more likely to rely on staying connected with their family and friends. Interestingly, almost no respondents reported using alcohol, kava, tobacco, or betel nut as a coping strategy. Figure 19. Coping strategies of households with children Other Recreation Seek guidance Help my kids Exercise Pray Read news Stay positive Stay connected with friends/family 0% 20% 40% 60% 80% Female-headed household Male-headed household Source: UNICEF mobile phone survey 35 In households with children, the most common coping strategy that respondents reported was using prayer with children (51.2 percent, CI: 38.6, 63.7). This data varied considerably based on the characteristics of the household. Respondents from male-headed households were more likely to report that children relied on prayer as their main coping strategy, whereas respondents from female-headed households were more likely to report that children relied on recreational activities and/or staying connected with friends and family (see Figure 20). Rural respondents were more likely to report that children were relying on recreational activities. Urban respondents were more likely to report that children were relying more on staying connected with friends and family. It must be emphasized that these results reflect the views of the adult respondent and not the views of children. Figure 20. Children’s coping strategies (for households with children only) Female-headed household Male-headed household Other Recreation Talk about worries Pray Spend time online Read news Stay positive Stay connected with friends/family 0% 20% 40% 60% Female-headed household Male-headed household Other Recreation Talk about worries Pray Spend time online Read news Stay positive Stay connected with friends/family 0% 20% 40% 60% Source: UNICEF mobile phone survey Urban Rural 36 4. Household Coping Strategies and Personal Coping Strategies In the UNICEF SIAS, around half of the respondents with children in their households, noticed a change in the behavior of children since the start of the COVID-19 pandemic. The most reported change in behavior of children was that children were worried (particularly that their family/friends would die or get sick from COVID-19). Other common reported changes in behavior from parents included children not completing homework, being more “clingy” than usual, and being angry. The main differences between households were that children in female-headed households were much more likely to be reported to be “clingy” and use internet and social media (see Figure 21). Like the results reported for the coping strategies of children, these results reflect the views of respondents and as such the differences reported may be due to differing views of respondents, rather than children’s behavior being systematically different in male-headed households versus female-headed households. Figure 21. Frequency of children’s behaviors that are more common than usual Other Not completing homework Clingy/More demand on parents More time in the internet/social media use Not sleeping or eating Sad Angry Worried 0% 20% 40% 60% Source: UNICEF mobile phone survey 37 5. Government response This section reports analysis from the World Bank HFPS on respondents’ awareness of the SIG’s Economic Stimulus Package (ESP)26. Those respondents who reported being aware of the ESP were asked about a range of the ESP elements that they may benefit from, including grants in the agricultural, fishing, or forestry sector, financial sector initiatives, and payouts from the national provident fund.the SIG quarantine facilities were also forced to close by the SIG. To prepare for the potential need for lockdowns in the event of a future outbreak, SIG has held “practice lockdowns” in Honiara which only last for a short time (1-3 days). 5.1 ESP Background The ESP was launched in May 2020 to support households and businesses during COVID-19, equivalent to approximately SBD309 million or 2.6% of GDP27 the ESP. The ESP comprises a range of initiatives, including cash grants to small and medium enterprises in agricultural, forestry, fishing, and tourism activities; a subsidy scheme for copra and/or cocoa producers; access to new loans through the Development Bank of Solomon Islands (DBSI), and temporary relief on existing loans; and access to payouts from the National Provident Fund (NPF) for members who have lost work due to COVID-19. It is important to note that since the time of the December 2020 HFPS, awareness of the ESP is likely to have grown as well as the number of households that have benefited. Future rounds of the HFPS will be able to assess ESP progress. 5.2 Awareness of the ESP Most respondents (60.5 percent, CI: 48.2, 71.6) had heard of the ESP at the time of the Round Two HFPS, but awareness varied by the location, wealth and sex of the respondent (see Figure 22). Respondents in rural areas were more likely to have not heard of the ESP (45.9 percent, CI: 31.0, 61.5), compared to Honiara (18.6 percent, CI: 11.3, 29.0). Only 16.8 percent (CI: 13.7, 20.3) of the top 20 percent of households had 26 Section 5 uses household level data. 27 https://solomons.gov.sb/solomon-islands-government-economic-stimulus-package-to-address-the-impacts-of-the-covid-19-pandemic/ 38 5. Government response not heard of the ESP compared to 55.9 percent (CI: 36.7, 73.6) of the poorest 40 percent of households. Female respondents were less likely to have heard of the ESP. Households with children (43.8 percent, CI: 31.7, 56.7) were more likely to have not heard of ESP than those without children (27.7 percent, CI: 9.7, 57.8). Figure 22. Awareness of ESP by sex, family structure, location, and well-being status Aware of ESP Not aware of ESP 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% urban 40% Quintile Quintile National Sex Location Wellbeing status Source: Round two of the World Bank high frequency mobile phone survey. 5.4 ESP Grant Applications and ESP Subsidies Of households conducting any of either agricultural, fishing, or forestry activities, 9.2 percent (CI: 5.8, 14.5) had reportedly applied to the ESP to receive grants28. The final outcome of some of these applications was not known at the time of the survey29. Households in the top 20 percent, in urban areas and with a male-headed households were more likely to have had applied for grants30. At the time of the survey, much less common were applications of households in the tourism sector. Nationally, 3.2 percent (CI: 2.4, 4.4)31 of households had reportedly applied for assistance through the tourism component of the ESP. The number of these applications that would be approved is unknown at the time of the survey. Future rounds of the HFPS should provide a fuller picture of the extent to which grants and subsidies have benefited households across the country. 28 Of those who have heard of the ESP, 18.3 percent (CI: 12.1, 26.6) of households conducting agriculture, fishing or forestry activities applied for grants. 29 Of applications made, around two thirds were rejected, with the rest being approved or were awaiting approval at the time of the survey. No data was collected on when the applications were made or when households expected to receive a decision. 30 Of those who have heard of the ESP, 27.4 percent (CI: 21.3, 34.5) of households in the top 20 percent, 22.8 percent (CI: 16.9, 30.1) in urban areas, and 18.8 percent (CI: 11.9, 28.6) with male head applied to the agriculture ESP. 31 Not reported in a table. Of those who have heard of the ESP, 5.3 percent (CI: 3.2, 7.9) applied to the tourism ESP. 39 Of the 6.5 percent of households producing copra and/or cocoa, just over a quarter had respondents that reported benefiting from the ESP subsidies for copra and/or cocoa (see Figure 23). For copra the SIG implemented a price subsidy of $0.50 per kg of copra and a freight subsidy of $28 per bag. For cocoa, SIG implemented a freight subsidy of $40 per bag. It is therefore possible that some respondents were not aware of their household benefiting from subsidies- for example some households may not have realized that reduced freight costs for cocoa were due to the subsidy scheme. Other sources of data (including administrative data) would be needed to determine the true extent to which subsidies have benefited copra and cocoa producers. Figure 23. Share of households that benefited from the copra and/or cocoa subsidy Does not farm copra/cocoa Did not receive subsidy Received subsidy Does not farm copra/cocoa Farms Did not receive subsidy 5% 93% copra/coca 7% Received subsidy 2% Source: Round two of the World Bank high frequency mobile phone survey. 5.6 Financial Sector Initiatives Of households who had heard of the ESP and had a loan with a formal lending institution, 8.6 percent (CI: 5.2 11.8) received financial relief through delaying loan repayments. Nationally, just 5.9 percent of households reported having an existing loan with a formal lending institution limiting the potential pool of households that could benefit from loan relief. This means that less than 1 percent of households reported receiving temporary relief on repayments from these loans as part of the ESP (see Figure 24). With such a small percentage of households benefiting, it is difficult to accurately characterize these households. Figure 24. Share of households that received temporary relief from loans Did not have an existing loan Did not receive relief Received temporary relief Did not have an existing loan 94% Did not receive relief 5% Had loan 6% Received temporary relief 1% Source: Round two of the World Bank high frequency mobile phone survey. 40 5. Government response The ESP helped some households cope with the economic consequences of COVID by increasing access to new loans (see Figure 25). Nationally, 5.9 percent (CI: 3.7, 9.5) of households took out a new loan from a financial institution since July 2020. Almost one in eight of these new loans were from the Development Bank of Solomon Islands (DBSI) as part of the ESP. New loans were more common among the richest 40 percent of households and households in urban areas. Households with children were around three times more likely than those without children to gain access to new loans. Figure 25. Share of households that have taken out a new loan since July 25% 20% 15% 10% 5% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one urban 40% Quintile Quintile child Sex of Family location Wellbeing status household structure head Source: Round two of the World Bank high frequency mobile phone survey 5.7 Payouts from National Provident Fund32 Of those households who had heard of the ESP, a least one member in 12.3 percent (CI: 6.6, 21.9) of households reported receiving a payout from the NPF as part of the ESP. The actual number of recipients may have been higher if there are multiple recipients in a single household or if respondents did not know about payouts to other household members. As the NPF only covers formal sector jobs, households in urban areas and households higher up in the welfare distribution would be more likely to benefit if they are more likely to have members who work in the formal sector work. Of households that heard of the ESP, 16.0 percent (CI: 11.7, 21.5) of households in Honiara received a NPF payout compared to 10.6 percent (CI: 3.5, 27.9) of rural households (see Figure 26). Households in the top 20 percent were more likely to receive a payout at 15.1 percent (CI: 12.1, 18.5), compared to only 6.8 percent (CI: 2.9, 15.3) of households in the third quintile. Households with children were more likely to benefit from NPF payouts (16.0 percent, CI: 8.5, 28.32) than those without children’s percent (3.0 percent, CI: 1.3, 6.7). 32 The NPF is a defined contribution fund established by the Solomon Islands National Provident Fund Act 1973. It is a financial institution to which members (employees) contribute a minimum of 5 percent and their employers contribute a 7.5 percent on behalf of their employee. 41 Figure 26. Share of households received a payout from the NPF, of households that have heard of the ESP 50% 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% male female No At least Honiara other rural Bottom 3rd 4th Top 20% children one child urban 40% Quintile Quintile Sex of household Family structure Location Wellbeing status head Source: Round two of the World Bank high frequency mobile phone survey. 42 6. Public Services This section analyses a range of critical public services, beginning with analysis of different categories of health care access and briefly reporting on a specific but highly relevant public health issue-vaccine hesitancy33. The section then reports data on water and sanitation, for drinking and hygiene. Lastly this section analyses education and schooling, which have been disrupted to some extent by lockdown restrictions. 6.1. Access to Health Care According to the UNICEF SIAS, almost half of the respondents who needed to access health care experienced difficulties in accessing health care services in the last week prior to being surveyed. Nationally, 43.8 percent (CI: 32,4, 56.0) of households needed health care services in the last week. Of those households that needed access, a large number 46.1 percent (CI: 27.8, 65.5) experienced difficulties with access (see Figure 27). reports the percentages of households who experienced difficulties in accessing health care, of those households who needed health care34. Figure 27. Difficulties to access health care services by households among those requiring care Had di culties accessing health care services, of those needing access 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH Head Family structure Location Wellbeing Status Source: UNICEF mobile phone survey. 33 Section 6 used household level data. 34 Of households that experienced difficulties in accessing healthcare, the majority confirmed that they were not able to access healthcare by providing a reason. However, for a significant proportion a reason was not provided so it’s not possible to provide a precise estimate on success in accessing healthcare. Questions about specific types of health care were asked in the World Bank HFPS, which reveals high rates of access. 43 Nationally, 24.8 percent (CI: 13.3, 41.4) of households indicated specifically needing urgent care in the month prior to the World Bank HFPS, of which 93.2 percent (CI: 83.0, 97.4) were able to access urgent care (see Figure 28). Below reports the shares of households requiring urgent care by group, as well as those that were able to access the needed care. It does not indicate whether or not these households experienced difficulties in accessing healthcare. Econometric analysis, reported in Table 10 in the appendix, showed those households which had higher shares of children under age 5 and those living in the third quintile and fourth quintile were less likely to be able to access needed urgent care, as well as those in Rennell-Bellona and Central provinces. The main reasons respondents were unable to access urgent care were the inability to travel (25.2 percent, CI: 7.5, 58.4), no medical personnel available (21.2 percent, CI: 6.6, 50.7), and a lack of money (52.4 percent, CI: 19.1, 83.7). Figure 28. Ability to access urgent care (total percentage of households requiring care indicated above bar) Able to access urgent care Unable to access urgent care 45% 42.0% 40% 35% 28.7% 29.3% 30.5% 30% 27.7% 26.3% 24.8% 22.7% 25% 19.2% 20% 15% 10.2% 10% 5% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex of HH Head Location Wellbeing Status Source: Round two of the World Bank high frequency mobile phone survey. Nationally, 15.2 percent (CI: 6.3, 32.3) of households indicated requiring access to routine care in the past month, of which 87.6 percent (CI: 63.7, 96.6) were able to access routine care. Routine care included tuberculosis care and family planning (see Figure 29). Below reports the shares of households requiring routine care by group, Figure 29. Ability to access routine care (total percentage of households requiring care indicated above bar) Able to access routine care Unable to access routine care 45% 37.8% 40% 35% 30% 22.9% 25% 18.4% 20% 15.2% 15.7% 15.8% 14.1% 13.6% 15% 7.8% 10% 4.6% 5% 0% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex of HH Head Location Wellbeing Status Source: Round two of the World Bank high frequency mobile phone survey. 44 6. Public Services as well as those that were able to access the needed care. Econometric analysis indicated that those living in Western and Isabel provinces had the most difficulty with access to needed routine care. See Table 10 in the appendix for econometric results. Of those not being able to access care, the main reasons reported were the inability to travel and a lack of medical personnel. Nationally, 22.6 percent (CI: 11.5, 39.5) of respondents indicated that at least one person in their household required preventative care in the past month, of which 95.6 percent (CI: 85.2, 98.8) were able to access preventive care. Preventative care includes childhood vaccines and pre-natal care (see Figure 30). Below reports the shares of households requiring preventative care by group, as well as those that were able to access the needed care. Econometric analysis (see Table 10 in the appendix) reveals those living in the bottom 40 percent were significantly less likely to be able to access needed preventative care, as were those households which had higher shares of children under age 2 and those over age 65. As reported, the main reasons for those who were not able to access care was that no medical personnel were available and the inability to travel. Figure 30. Ability to access preventative care (total percentage of households requiring care indicated above bar) Able to access preventative care Unable to access preventative care 45% 41.5% 40% 35% 30.9% 29.0% 28.6% 30% 25.0% 25% 22.6% 23.2% 19.2% 20.1% 20% 16.4% 16.3% 15% 10% 7.4% 5% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH Head Family structure Location Wellbeing Status Source: Round two of the World Bank high frequency mobile phone survey. Of those needing to buy medicine in the last week prior to the UNICEF SIAS, 43.4% (CI: 20.5, 69.5) experienced difficulties. Almost one fifth of respondents needed to buy medicine in the last week prior to the UNICEF SIAS (18.4 percent, CI: 10.7, 29.6). Of these households, Figure 31 reports the percentage of households that had difficulties buying medicine, by group. Female-headed households were more likely to have had difficulties (81.7 percent, CI: 60.9, 92.8), compared to male-headed households (34.9 percent, CI: 13.0, 65.8). Though not a statistically significant difference, 39.9 percent (CI: 17.4, 67.6) of households with children had difficulties in buying medicine compared to 70.6 percent (CI: 23.7, 94.9) in households without children. While those in Honiara were less likely to have experienced difficulties, the difference is not statistically significant. 45 Figure 31. Difficulties in buying medicine by households among those needing medicine Had di culties accessing medicine, of those needing access 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH Head Family structure Location Wellbeing Status Source: UNICEF mobile phone survey 6.2. COVID-19 Vaccination35 Vaccine hesitancy is a widespread and significant issue, according to data from the UNICEF SIAS. Around half of respondents (47.4 percent, CI: 35.9, 59.1) would get vaccinated if they were offered an approved vaccine at no cost, while 28.9 percent (CI: 19.1, 41.2) would not, and about one quarter (23.7 percent; CI: 16.3, 33.2) were unsure. Willingness to get vaccinated did not vary significantly between respondents in rural and urban areas36. In terms of the family structure, 59.6 percent (CI: 31.8, 82.4) of households without children were willing to get vaccinated, compared to 44.9 percent (CI: 32.6, 57.9) of those with children. Respondents who did not want to take a vaccine for COVID-19, or were unsure, were asked to provide their main concern. Concerns included: they are worried about side-effects, they do not trust vaccines in general, they don’t think it will work, they do not believe they are at risk of COVID-19; and a vaccine is against their religion; the health facility is too far or hard to get to; they are worried about getting COVID-19 at the health facility (see Figure 32). The most common main concern was due to worry about side-effects at 30.1 percent (CI:18.9, 46.1). The two other most common concerns were mistrust in vaccines in general and respondents not thinking the vaccine will be safe. The next round of the World Bank HFPS asks a broader range of vaccine questions that will allow for deeper analysis. 35 It is a financial institution to which members (employees) contribute a minimum of 5 percent and their employers contribute 7.5 percent on behalf of the employee. 36 Willingness to get vaccinated also did not vary greatly between in male- and female-headed households (48.1 and 44.0 percent respectively), but as the question was asked of the respondent, this result is not insightful. The sex of respondents in the UNICEF survey was not recorded. 46 6. Public Services Figure 32. Main concern for not wanting to take COVID-19 vaccination Distance to health facility Don't trust vaccines in general Not at risk of COVID-19 Worried about side e ects I dont think it will be safe Don't think it will work Other 0% 10% 20% 30% 40% 50% Source: UNICEF mobile phone survey Note: Categories that constitute less than 1 percent of responses have been combined into “other” 6.3. Water and Sanitation Around one fifth (22.0 percent, CI: 14.0, 32.7) of households reported having insufficient drinking water in the last week. For these households, the main reasons households were unable to access drinking water was due to inadequate supply (53.8 percent, CI: 31.4, 74.8) and a lack of affordability (22.4 percent, CI: 7.4, 51.2). More than half of female-headed households cited a lack of affordability (51.7 percent, CI: 11.4, 89.9) as the main reason they could not access drinking water compared to just over 12.2 percent (CI: 4.9, 27.5) of male-headed households (see Figure 33). Inadequate supply was a major issue for most households, particularly those without children (88.9 percent, CI: 53.5, 98.2) and those in rural areas (59.4 percent, CI: 30.6, 83.0). Figure 33. Main reason households were unable to access water for drinking Supply not available Supply reduced Unable to access shared sources Unable to a ord water Other 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH head Family structure Location Well-being status Source: UNICEF mobile phone survey 47 Almost three-quarters (73.2 percent, CI: 59.2, 83.7) of households reported adequate access to water and soap for hand-washing. In urban areas 87.7 percent (CI: 71.0, 95.4) of female-headed households reported they had sufficient access to soap and water for hand-washing compared to 86.1 percent of male-headed households (CI: 79.0, 91.1) (see Figure 34). In contrast, 48.0 percent (CI: 19.2, 78.2) of female-headed households in rural areas had sufficient access to soap and water, compared to 74.0 percent of male-headed households (CI: 54.2, 87.2). In female-headed households with children, 52.8 percent (CI: 24.7, 79.2) reported adequate access which is lower than for male-headed households with children (75.2 percent, CI: 56.6, 87.5). The main reason for the lack of access across all households who reported inadequate access was affordability– 92.7 percent (CI: 77.7,97.9) of households who didn’t have adequate access stated they were unable to afford soap, while only 1.75 percent (CI: 0.4, 6.9) of households stated that soap was not available at their local market. Figure 34. Share of households having sufficient soap and water to wash hands in the last week Share of households having su cient soap and water 100% 80% 60% 40% 20% 0% Male head Female Male head Female Male head Female Male head Female head head head head Rural Urban No children At leat one child National Location Family structure Source: UNICEF mobile phone survey. Around half of respondents reported that members in their household wash their hands using water from a tap/sink in their dwelling (10.6 percent, CI: 6.5,16.8 ) or a tap/sink in their yard (40.1 percent, CI: 29.7, 51.4). Almost a third (30.9 percent, CI: 21.3, 42.4) of households relied on a bucket to wash their hands and just 3.4 percent (CI: 0.6, 16.1) of households reported not having a place to wash hands. Using a tap/ sink for hand-washing was far more common in urban areas (68.8 percent, CI: 57.2, 78.4) compared to rural areas (45.8 percent, CI: 32.1, 60.1) and more common in households without children (74.9 percent, CI: 50.7, 89.7) and female-headed households (65.5 percent, CI: 36.3, 86.4, see Figure 35). Almost no male-headed households (0.5 percent. CI: 0.1, 2.6) reported having no hand-washing facility at home, whereas 16.3 percent (CI: 2.5, 59.6) of female-headed households reported that they do not have a hand-washing facility at home. 48 6. Public Services Figure 35. Type of hand-washing facility Tap/sink in dwelling Tap/sink in yard Bucket No handwashing Other 100% 80% 60% 40% 20% 0% Male Female No At least Honiara Other Rural Bottom 3rd 4th Top 20% children one child Urban 40% Quintile Quintile National Sex of HH head Family structure Location Well-being status Source: UNICEF mobile phone survey 6.4. Education and Schooling Households with children were asked questions specifically about these children37. Households with at least one child below 15 years of age constituted 85.6 percent of HFPS sample and 78.2 percent of the UNICEF SIAS sample (Appendix 2, Table 2). Households with school aged children (6-14 years38) accounted for 53.5 percent of the UNICEF SIAS sample (noting the two groups of households are not mutually exclusive). The government COVID-19 containment measures and response to the pandemic may have caused a range of disruptions to education, though schools were only shutdown for the initial period of March- May 2020. This section draws on the UNICEF SIAS data to explore these issues. The location at which children spent most of the time during lockdown was at home39. This was true for both households with children aged 0-5 years (82.4 percent, CI: 61.5, 93.2) and households with school aged children (76.9 percent, CI: 60.6, 87.8). For households with school aged children, then next most commonly location in which children spent most of their time during lockdowns was in school, at 6.7 percent (CI: 3.1, 13.6). Of households with children age 0-5 years, 60.4 percent (CI: 42.6, 75.9) expressed concern with the arrangements for children during lockdowns. This was the same for households with children age 6-14 years, of which 60.3 percent (CI: 44.4, 74.3) expressed concern with the arrangements. For households with children 0-5 years old the most commonly cited concern40 about the arrangements was regarding children’s safety (55.9 percent, CI: 36.9, 73.3). This was also the most 37 The statistics reported in this section for age groups (0-5 and 6-14) are calculated using weights so that the sample of households is representative of the population of households with children in those age groups. Households without children are excluded from the analysis in this section. 38 For the purposes of analytical clarity, “school aged children” always refers to the age group 6-14 in this report. 39 Lockdown” is broadly defined and refers to either lockdowns or state of emergency measures that may close schools. The survey question regarding where children spent most of their time during lockdown had the following response options: inside the home, outside going farming or chores, someone’s work-place, or another household. 40 Household were asked about the reason for their concern, with the following response options - “worried about safety”, “worried about learning outcomes”, “having to pay money for the arrangement”, “worried about children’s health”, and “worried about (the respondent’s) productivity” 49 commonly cited concern for households with children 6-14 years old at 43.8 percent (CI: 27.9, 61.0). The next most commonly cited concern regarding children staying at home for households with school aged children was the impact on children’s learning outcomes (26.8 percent, CI: 13.0, 47.1). A majority of respondents (60.2 percent, CI: 42.2, 75.8) from households with school-aged children said that all or some children in their household completed some education at home during the school shutdown. There was no significant difference across location or the sex of the household head. The most common form of education at home was that adults helped with the children’s learning (77.3 percent; CI: 59.1, 89.0), followed by independent reading or work (60.2 percent; CI: 39.8, 77.7), and private tutoring (31.9 percent, CI: 16.1, 53.4). Respondents in urban areas were more likely to listen to radio programs41 (14.6 percent (CI: 9.6, 22.5) compared to 1.8 percent (CI: 4.3, 7.2) in rural areas. Respondents in female-headed households were more likely to report “independent reading” as a form of education (82.3 percent, CI: 48.4, 95.9) than male-headed households (55.7 percent, CI: 33.1, 76.1). Respondents in female-headed households were less likely to report that “adults helped with learning” (33.1 percent, CI: 9.4, 70.4) than male-headed households (86.5 percent, CI: 76.3, 92.8). Respondents were in female-headed households less likely to report that “children received private tutoring” (8.6 percent, CI: 2.9, 23.1), than male-headed households (36.8 percent, CI: 18.0, 60.6). A minority of households with school-aged children (9.4 percent, CI: 5.4, 16.0) reported that children in their household received some form of remote learning support. Respondents in urban areas were more likely to have received support (16.1 percent, CI: 11.3, 22.3) than respondents in rural areas (7.6 percent, CI: 3.3, 16.3). The form of information and services that were available by school for remote learning support was mainly paper based42 (84.8 percent; CI: 65.3, 94.3). When asked about the way the respondent supported their children with remote learning, the most common answer was they spent more time with children (72.3 percent, CI: 56.6, 83.9), followed by private tutoring (30.2 percent, CI: 18.0, 46.0). Private tutoring was more commonly reported in urban areas (38.7 percent, CI: 28.8, 49.6) than in rural areas (27.8 percent, CI: 13.6, 48.4). Respondents in female-headed households were less likely to report that they spent more time with their children (65.0 percent, CI: 30.1, 88.9) than male-headed households (73.7 percent, CI: 56.0, 86.0). Also, female-headed households were less likely to report private tutoring (33.6 percent, CI: 10.4, 68.6) compared to male-headed households (29.5 percent, CI: 16.3, 47.3). Most respondents answered that the most important educational requirement that should be provided for home-based schooling was the availability of paper materials (65.7 percent, CI: 48.4, 79.6), followed by guidance for parents (41.1 percent CI: 25.9, 58.1). See Figure 36 that reports educational requirements. 41 It should be noted that while radio programs were considered a form of education by the respondent, there isn’t data collected on whether these radio programs were run by schools or were specifically designed to be educational 42 No information was collected on how paper based learning supports were distributed to children 50 6. Public Services Figure 36. Key educational requirements for home based schooling (multi-select) Online teaching/Tutorials Availability of materials and equipment for online platform learning Availability of paper-based materials Guidance for parents Remote support from teachers Support for children with special needs Other 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Source: UNICEF mobile phone survey The most common challenge for children with home learning was lack of access to textbooks at home (67.5 percent, CI: 49.6, 81.4), followed by availability of worksheets, (46.7 percent, CI: 30.3, 63.8), and that the worksheets were boring (43.3 percent, CI: 28.1, 59.8)43. A higher proportion of respondents from urban areas reported that parents cannot provide sufficient guidance as was normally provided by teachers (57.2 percent, CI: 46.9, 66.9) compared to rural areas (33.4 percent, CI: 17.5, 54.2) (see Figure 37). Figure 37. Main Challenges for home based schooling Urban Rural They do not have worksheets They do not have textbooks at home Children have to help household with other priorities (housework, childcare, etc.) Parents/caregivers cannot provide su cient guidance or support Children find reading/worksheets boring/less interesting than what they do in school Other 0% 10% 20% 30% 40% 50% 60% 70% 80% Source: UNICEF mobile phone survey 43 The national estimates are not reported in the chart. 51 Around four fifths (83.4 percent; CI: 60.5, 94.3) of respondents from households with school-aged children expressed concern about a child’s school performance and learning during the lockdown. Respondents in female-headed households and from urban areas were more likely to express concern about their children’s performance and learning. The most cited concern across all household types was that children may lose interest in school or learning (57.3.1 percent, CI: 40.3, 72.7), followed by children may not pass exams or school this year (23.5 percent, CI: 12.2, 40.2). While respondents in male-headed households were more concerned that children may lose interest in school (64.1 percent; CI: 47.2, 78.1), respondents in female-headed households were more concerned about children not passing their exams for school this year (58.2 percent, CI: 23.6, 86.2). The majority of respondents from households with school-aged children said children usually attended school before the lockdown (87.8 percent; CI: 75.0, 94.6). The percentage was higher among respondents in urban areas than in rural areas (95.8 percent; CI: 92.4, 97.7, compared to 85.7 percent; CI: 69.3, 94.1), and also higher among respondents in male-headed households (90.0 percent, CI: 75.9, 96.3) than in female-headed households (77.5 percent, CI: 40.4, 94.6). A majority of parents reported that most schools were in their community were closed during lockdown (80.2 percent; CI: 64.4). The percentage was higher for households who lived in urban areas (85.2 percent; CI: 74.6, 91.8) than in rural areas (78.8 percent; CI: 58.8,90.7) but not by a significant margin. The difference was relatively large between female-headed households and male-headed households (97.5 percent; CI: 92.1, 99.3) and 76.4 percent (CI: 58.4, 88.2 respectively). Most respondents reported that children had access to items for their education and entertainment at home. The vast majority of respondents from households with children aged 0-5 years mentioned that children had access to toys (66.4 percent; CI: 48.0, 81.0), and books (55.1 percent; CI: 38.5, 70.6, see Figure 38). Access to these items was different among children aged 6-14 years, with two-thirds of respondents mentioned access to a book (68.1 percent, CI: 53.7, 79.7) and fewer (42.6 percent CI: 29.0, 57.4) mentioned access to a smartphone. In both age groups, access to all items was higher in urban areas than in rural areas, except for access to radio. Figure 38. Children’s access in the household Toys Children book Smart phone Age 6-14 Radio Age 0-5 Computer Television 0% 20% 40% 60% 80% Source: UNICEF mobile phone survey 52 7. Public Trust & Community Security This section captures people’s perceptions of public trust and community security but does not directly measure any objective indictors of these issues (e.g., official statistics on actual reported crime)44. Respondents were asked if they think that there have been changes to trust and social relations with members of their community. Limited comparisons can be made between the findings reported here and the findings from the first World Bank HFPS (June 2020), in which a broader set of questions were asked. In the first round of the HFPS, most respondents said that the public trust and safety within the community had remained the same since January, though there were still some causes for concern. For example, the first round of the HFPS found that in rural areas, women were more likely to say that things had deteriorated due to domestic abuse. Given that pre-crisis levels of gender-based violence were among the highest in the world, increased domestic abuse since the COVID-19 restrictions began, could be an area of ongoing concern. However, there is the difficulty measuring such issues using phone surveys. It is important to note that the data does not reflect any developments in relation to public trust and community security that occurred in 2021. 7.1. Public Trust Community trust and social relations remain resilient, with little evidence of any perceived deterioration between June and December 2020. Overall, 65.5 percent (CI: 49.3, 78.8) of respondents indicated that trust and social relations within the community had remained the same in the six months prior to the survey, with 22.0 percent (CI: 12.8, 35.1) saying relations had improved and 12.5 percent (CI: 4.6, 29.6) indicating that they believed they had deteriorated. Figure 39 below reports the findings across groups. The share reporting that the situation had improved in this period was significantly lower for households in the bottom 40 percent at 13.3 percent (CI: 2.6, 47.2), compared to 33.2 percent (CI: 24.8, 42.8) in the fourth quintile. Since baseline levels are unknown it is therefore not possible to conclude the relative levels of satisfaction with different problems in different locations, only the changes since the pre-crisis period45. Similarly, the majority of respondents in the June 2020 survey reported that trust and social relations within the community stayed the same since the start of 2020. 44 Section 7 uses household level data. 45 For example, an area experiencing a high level of violence within the community may have seen a relative improvement since the crisis, but the overall levels could still be higher than a peaceful area that has seen a deterioration. 53 Figure 39. Change in trust and social relations since June/July, by sex of respondent, location, and wellbeing status Improved Stayed the same Deteriorated 80% 60% 40% 20% 0% -20% -40% -60% Male Female Honiara Other Rural Bottom 3rd 4th Top 20% Urban 40% Quintile Quintile National Sex of Respondent Location Wellbeing Status Source: Round two of the World Bank high frequency mobile phone survey 7.2. Community Security Issues For issues related to sensitive topics, a series of questions about the change in levels within the community were asked. The topics included in this section were theft, damage to property, physical assault, verbal abuse, alcohol and drug abuse, intimidation by police, violence by police, land disputes, and domestic abuse. The survey methodology literature has shown that respondents are more likely to misrepresent the truth if asked sensitive questions directly, either out of embarrassment or fear of retaliation by an aggressor. Asking about the community as a whole is therefore a recommended method to obtain high quality information about the change in these indicators without potentially endangering respondents. Higher percentages of respondents in urban areas reported improvements in community security issues compared to rural areas. Figure 40 and Figure 41 provide a breakdown by security issue, for urban and rural areas, respectively. Urban respondents were significantly more likely compared to rural respondents to say that things had improved since June/July 2020 with regard to theft (46.4 percent (CI: 38.3, 54.7) compared to 39.2 percent (CI: 17.1, 66.8) in rural areas. This was also the case for property damage, at 52.2 percent (CI: 44.2, 60.0) compared to 33.5 percent, (CI: 14.6, 59.8), and police intimidation (43.3 percent, CI: 36.1, 50.9, compared to 31.1 percent, CI: 11.5, 61.3). On the latter, given the extended state of emergency, it is noteworthy that people’s perception on police force remains relatively unchanged or improved- these findings seem to counter concern that restrictions and enforcement led to abuse of power. In both urban areas and rural areas, significant shares of respondents reported an improvement in domestic abuse, while smaller but substantial shares reported a deterioration. A higher percentage of respondents in rural areas (54.3 percent, CI: 27.4, 78.9) reported an improvement, compared to urban areas (40.6 percent, CI: 33.6, 48.0). Roughly equal shares of respondents in urban and rural areas reported that 54 7. Public Trust & Community Security things deteriorated with regard and domestic abuse (16.7 percent compared to 16.1 percent). Overall, given the subjective nature of the data and the lack of consensus, it is difficult to say whether or not domestic violence actually did increase or decrease. On a related note, in both rural areas and urban areas, alcohol and drug abuse were the most commonly reported issue to have gotten worse. This could suggest an increased risk of domestic violence in the future. Gender differences in views of security issues follow different patterns in urban and rural areas, with women in rural areas being more likely than women in urban areas to report a deterioration in the situation regarding domestic abuse. Econometric analysis provides further disaggregation of the general trends seen in urban and rural areas (Table 11 and Table 12, Appendix). In urban areas, women were significantly more likely than men to indicate that the situation had improved in regard to damage to property, and regarding theft, and land disputes. In contrast, in rural areas, women were less likely to report an improvement in damage to property and police intimidation. In rural areas, women were significantly more likely than men to say that the situation with domestic violence had deteriorated. This finding corresponds to the June 2020 round of the survey. In contrast, women in urban areas were significantly less likely than men to say that the situation with domestic violence had deteriorated. However, these differences need to be considered in relation to the findings in the previous paragraph. The bottom 40 percent and top 20 percent had different views regarding changes in some security issues, as revealed by econometric analysis (Table 11, Appendix). The bottom 40 percent in urban areas were less likely than the top 20 percent to say Figure 40.Change in community safety situation since June/July 2020 [urban] Improved Stayed the same Deteriorated Women Age 18 -25 Other urban Theft third quintile Age 46 -55 Age 56 -65 Age 56 -65 Damage to Property No school No education Primary education Vocational education Age 56-65 Age 56 -65; Age 66+ Alcohol and Drug Abuse 3 rd quintile Bottom 40% Women Age 66+ No education Intimidation by Police Primary education Bottom 40% Other urban Age 18 -25 No education Primary Land Disputes Primary education education Vocational education Other urban Other urban Bottom 40% Domestic Abuse Age 66+ Other urban Age 26 -35 third quintile 100% 80% 60% 40% 20% 0% 20% 40% 60% 80% 100% Source: Round two of the World Bank high frequency mobile phone survey Note: Calculations based on the December 2020 round of the high frequency mobile phone survey. Results show weighted percentage of respondents indicated the situation within the community had gotten better, remained the same, or gotten worse. Characteristics listed on the “deteriorated” side of the graph were statistically significantly associated with indicating things had gotten worse. Characteristics listed on the “improved” side of the graph were statistically significantly associated with indicating things had gotten better. 55 that things improved regarding community trust, alcohol and drug abuse and land disputes. Correspondingly, in terms of a deterioration, those in the bottom 40 percent in urban areas were more likely to say the situation in regard to community trust and alcohol and drug had deteriorated. The results indicate some level of social disenfranchisement due to the restrictions or other implications of the pandemic, that disproportionately impact the bottom 40 in urban areas. Figure 41. Change in community safety situation since June/July 2020 [rural] Improved Stayed the same Deteriorated Women Age 18 -25 Noeducation Tertiary education Women Age 46-55; Age 66+ third quintile Theft No education Women Primary education Age 26-35; Age third quintile 66+ Age 18 -25; Age 46 -55 Primary Damage to Property Age 56 -65; Age 66+ education Vocational education Women fourth quintile Age 46-55; Age 56-65; Age 66+ Women Alcohol and Drug Abuse No education Age46-55; Age 56 -65; third quintile Age 66+ third quintile Women Age46-55 Age 18-25; Age 56-65 Age 56 -65; Age 66+ Intimidation by Police Primary education No education Bottom 40% Tertiary education third quintile Age 18-25; Age 26-35; Age 66+ Bottom 40% Women Land Disputes third quintile Age18-25 Bottom 40% Age18- Women third quintile 25 Age 18 -25; Age 26 -35 No Domestic Abuse No education education Primary education Bottom 40% third 80% 60% 40% 20% 0% 20% 40% 60% 80% quintile fourth Source: Round Two of the World Bank High Frequency Mobile Phone Survey Note: Calculations based on the December 2020 round of the High Frequency Mobile Phone Survey. Results show weighted percentages of respondents indicate the situation within the community had gotten better, remained the same, or gotten worse. Characteristics listed on the “deteriorated” side of the graph were statistically significantly associated with indicating things had gotten worse. Characteristics listed on the “improved” side of the graph were statistically significantly associated with indicating things had gotten better. Significant shares of respondents indicated deterioration in some security issues such as alcohol and drug abuse, and theft, across both the first half and second half of 2020. Some analysis over time is possible based on a sample that participated in both Round One and Round Two46. For Round One data, the reference period was January to June 2020, compared to the Round Two reference period of June to December 2020 (see Figure 42) below shows heat maps for perceived changes for several aspects of public trust and security across the two rounds. For non-violent 46 For heat map analysis, only those households appearing in both Rounds could be used – 1,048 total observations. A separate set of weights was therefore calculated using the same procedures described in the appendix to analyze this subsample. These results should be interpreted with caution due to the highly non-random nature of attrition between Rounds. 56 7. Public Trust & Community Security crimes and security issues47, between 20 to 40 percent of respondents indicated that the situation had deteriorated between both January and June as well as between June and December. Much smaller shares, generally 6 percent or less, indicated continued improvement. The next round of HFPS will reveal whether there have been significant changes in 2021. Figure 42. Heatmaps for changes in public trust and security Source: Rounds One and Two of the World Bank High Frequency Mobile Phone Survey. 47 Theft, damage to property, land disputes, and drug and alcohol abuse. 57 8. Conclusions and Policy Considerations This section first summariness key results of the analysis that may have policy implications. Based on these results, some broad areas for policy consideration are outlined. These areas include economic recovery and equity, protection of the vulnerable, public services, and public information regarding government policies. 8.1. Conclusions The economic impacts at the onset of the COVID-19 pandemic were widespread. The partial recovery in the second half of 2020 has been uneven, which if continued into 2021, would threaten rising inequality. By December 2020, employment had recovered only slightly from mid-2020 and had not returned to pre-pandemic levels. While formal sector employment and informal sector employment both declined from January to June 2020, in the second half of 2020 there was a stronger recovery for those in the formal sector and urban areas. Employment recovery was also stronger for households in Honiara and households in the third quintile while no recovery was experienced by those in the bottom 40 percent. Many agricultural households anticipate falling incomes and were more pessimistic in December 2020 than in June 2020. These results suggests that rural households, agricultural households and the informal sector are lagging in the recovery. As a consequence, inequality between sectors may have started to rise in 2020. The shape and speed of recovery will need to be monitored with both macro-economic data, and household-level data from future HFPS rounds. Vulnerability to poverty is likely also growing as most households are using medium-term coping strategies, such as drawing down savings, with informal safety nets providing less assistance to households. Possibly due to the persistent nature of the economic slowdown, households appear to have increased medium-term coping strategies in the second half of 2020, while continuing to employ short-term strategies such as reducing consumption. These strategies include seeking additional income generating activities, delay making re-payments, and drawing down on savings. Previously vulnerable households may fall into poverty, with reduced chances of recovery, given the reduction in their wealth, increased debt, and more limited 58 8. Conclusions and Policy Considerations income earning opportunities. There may also be a mental health toll of the pandemic as many respondents report that children are worried about the prospect of family and friends falling ill with COVID-19. Evidence supports SIG’s ESP benefiting households in 2020, but awareness of the ESP was low at the time of the Round Two HFPS. Awareness varied across groups with respondents in rural areas and those in the bottom 40 percent being less likely to have heard of the ESP. Similarly, applications to ESP grants varied by group with households in urban areas and the top 20 percent more likely to have applied. That a significant share of applications were awaiting decisions, may indicate an implementation challenge at that time (though one that may long since have been resolved). Urban households and households in the top 20 percent were also more likely to have reported benefiting from NPF payouts. Further HFPS data collection will be needed to determine if awareness of the ESP has improved and to measure the extent to which households across the country have benefited. In terms of public service provision, Solomon Islands is yet to be tested by any substantial outbreak of COVID-19 but faces challenges such as vaccine hesitancy, in preparing for a future with COVID-19. Health care service access has not been severely hampered by the crisis. The data was collected before Solomon Islands vaccination program ramped up, which may have diverted scarce health resources. In relation to vaccinations for COVID-19, a key constraint will be the half of the population that are either unsure or unwilling to be vaccinated, mostly because they are concerned about side-effects but also due to safety concerns and a general mistrust of vaccines. Another challenge for public service provision is that a significant number of households with children chose to reduce the number of these children attending school as a coping strategy, which would have implications for the long-term development of Solomon Islands. Lastly, of some concern is that some households do not have sufficient access to drinking water, mainly due to inadequate supply and a lack of affordability. Overall, there is not a widespread perception of deterioration in community trust and social relations, though perceptions vary across rural and urban areas, by gender and for the bottom 40. In rural areas for example, women are more likely to report a deterioration in domestic violence. While the results in relation to domestic violence are mixed, this should be considered more broadly, against a backdrop of extremely high pre-crisis levels of gender-based violence. On top of pre-existing anxieties regarding community and domestic security, adults and children are also worried about the prospect of COVID-19 entering Solomon Islands. 59 8.2 Policy considerations Given the limited and uneven nature of the recovery in employment and incomes, equitable growth should remain a priority. The Government of Solomon Islands could potentially promote employment growth through business support (as is occurring through the ESP) and labor-intensive public investment projects. Expanding the coverage of formal safety nets would greatly assist the households most in need given the coping strategies they use may not be sustainable over the long term. Government led social protection will be of greater relative importance the longer the crises endures. It is important to note that any stimulus measures and expansion in safety nets will have to be balanced against fiscal constraints- if the SIG is overextended it will be difficult to sustain economic support. The risk is slower growth, higher poverty, and increased inequality. The next HFPS should reveal whether the nascent recovery has deepened and broadened. Preparedness for a potential future outbreak of COVID-19 in Solomon Islands may be informed by data on access to basic services including health care, WASH, welfare services and education. One critical area for consideration regards the barriers and constraints to children staying enrolled in school and attending school. Going forward, it will be important to monitor school enrolments and attendance and consider ways for schools to support remote learning whenever remote learning is necessary. In terms of public health, policy makers could seek ways of expanding the supply of WASH services including clean drinking water, or make clean water more affordable. It is also important that health services are prepared to respond to any covid outbreak in a way that does not jeopardize the currently high rates of access to health care for other reasons. Notwithstanding vaccine hesitancy, the expanding distribution of vaccines in Solomon Islands will be critical in protecting the population from any future outbreak of COVID-19. Consideration of policies to protect those who are socially vulnerable will be important for the future. Expanded outreach and welfare services to address domestic violence (including violence against children) appear warranted, particularly in rural areas, where women are more likely to report a deterioration in domestic violence. Working with church and community leaders to mitigate detrimental behaviors such as drug and alcohol abuse (for which over a third of respondents perceive to be getting worse), is another option. Measures that reduce the economic insecurities which may underpin increased drug and alcohol abuse, including various safety nets could have both economic and social benefits. Given the stress placed on households of the recession, and the high levels of reported worry and anxiety, services that support mental health are another important consideration. 60 8. Conclusions and Policy Considerations An important consideration for several areas of public policy is the provision of, and targeting of, information to the public. Given that many households had not heard of the ESP at the time of the Round Two HFPS, greater effort may be needed in disseminating appropriate public information for multifaceted policy packages like the ESP. Given that awareness varied across groups, the targeting of public information is also likely important. Better communication of targeted information regarding COVID-19 vaccines, may be one potential pathway to reducing vaccine hesitancy. This approach could include information on vaccine side-effects and safety, as well as information on how vaccines can be accessed including collaborating with communities and religious organizations, as the roll out of the vaccination program expands. However, as identified from Round One, given the high prevalence of informal information sharing, managing information around COVID-19 is challenging. 61 Appendix 1: Technical A1.1. HFPS Instrument Design The survey instrument for the World Bank was designed by the project team based on the Round One questionnaire and the advice of the World Bank’s COVID-19 questionnaire working group. The team also consulted with staff in the Sydney and Port Moresby offices, in particular the Health and Agriculture teams, as well as received feedback from other development partners. As in Round One, the length of the survey was targeted at 15 minutes of call time. The questionnaire included the following modules: Basic Information, Employment and Income Loss, Coping Strategies, Health, Public Trust and Security, Assets, and Wellbeing. As in the first round, if the respondent was not the head of household, the Employment and Income Loss section included additional questions asking about the head of the household specifically. The survey instrument for Round Two consisted of 266 questions, but only a sub-set of questions were asked to each household, with the sections on Health, Public Trust and Security, and Assets and Wellbeing being asked only to a subset of households. For households that participated in both Round One and Round Two, the asset data collected in Round One was used to estimate which wealth decile they are in. A1.2. Sampling The target geographic distribution for the survey was based on the population distribution across provinces from the preliminary 2019 census results. According to the population census, Honiara constituted almost one quarter (18.0 percent) of the total population. However, as in Round One, Honiara was over-represented in the World Bank HFPS (constituting 32.8 percent of the survey sample). All other provinces were deemed under-represented, with the largest differences being for Makira-Ulawa, which represented 3.9 percent of the survey sample compared to 7.2 percent of the population in the census, and Guadalcanal, which represented 14.3 percent of the survey sample compared to 21.4 percent of the population in the census. Urban areas constituted almost half (49.2 percent) of the survey sample, compared to a quarter 62 Appendix 1: Technical (25.6 percent) of the census. Compensating factors for these differences were developed and included in the re-weighting calculations (see Figure 43). See section A.1.3 below for further information. Figure 43. World Bank HFPS Sample population distribution compared to 2019 census census survey 35% 30% 25% 20% 15% 10% 5% 0% ul al el ra u rn al ta na a ot tr ise ab ia w an ai te en llo m on la al Is ho es lc Te -U C Be M H da W C ira l- ua el ak G nn M Re The UNICEF SIAS successfully interviewed 1,530 households from the 2,881 households in the World Bank HFPS Round Two sample. This means that the attrition rate from the Round Two survey to the UNICEF SIAS was 46.9 percent. Some of the World Bank sample and the UNICEF SIAS sample characteristics, are compared in table 2. These include gender of household head, family structure, location, and wellbeing status of each sample A1.3. Weighting A1.3.1 World Bank Survey In Round One, there were two sets of primary weights calculated: those at the individual level and those at the household level. For the second round, it is possible to calculate four sets of primary weights: cross section weights at the household level, panel weights at the household level, cross section weights at the individual level, and panel weights at the individual level. Given the high levels of attrition, the panel weights were only used for selected analysis at the household level (as noted in the main text) and the majority of the analysis relied on cross sectional weights. The data collected from the mobile phone surveys differed substantially from the characteristics of the population as a whole (see Figure 44 below). The data required re-weighting. The process for reweighting was similar to that used in Round One (see 63 technical appendix to that report for full details). As shown in figure 45, re-weighting was generally successful in adjusting for differences in observable characteristics across the two rounds of mobile phone surveys as compared to the 2015 DHS. Figure 46 reports the distribution of the samples across the DHS wealth deciles. The Round Two sample is not as overrepresented in the top two quintiles as the Round One sample. Figure 44. Distribution of key characteristics prior to re-weighting DHS HFPS R1 HFPS R2 75% 50% 25% 0% 65+ 36 - 45 46 - 55 56 - 65 male 26 - 35 18 - 25 female secondary vocational primary tertiary no education sex age eduation Source: 2015 DHS and rounds one and two of the World Bank High Frequency Mobile Phone Survey Figure 45. Distribution of key characteristics after re-weighting DHS HFPS R1 HFPS R2 75% 50% 25% 0% 65+ 36 - 45 46 - 55 56 - 65 26 - 35 male 18 - 25 female primary tertiary secondary vocational no education sex age eduation Source: 2015 DHS and rounds one and two of the World Bank high frequency mobile phone survey. 64 Appendix 1: Technical Figure 46. Distribution of respondents across DHS deciles 40% 30% 20% 10% 0% 1 2 3 4 5 6 7 8 9 10 HFPS R1 HFPS R2 DHS Source: 2015 DHS and rounds one and two of the World Bank high frequency mobile phone survey. 65 Appendix 2: Tables Table 2. Number of respondents, by gender of household head, family structure, location, and wellbeing status (household level) World Bank UNICEF Number of Shares Number of Shares observations (percent) observations (percent) Total 2,882 100.0 1,530 100.0 Male-headed 2,365 82.1 1,267 82.8 Gender of household head Female-headed 517 17.9 263 17.2 No children 415 14.4 333 21.8 Family structure 47 At least one child 2,467 85.6 1,197 78.2 Honiara 945 32.8 514 33.6 Location Other urban 472 16.4 267 17.5 Rural 1,465 50.8 749 49.0 Bottom 40% 141 4.9 68 4.4 Third quintile 321 11.1 175 11.4 Wellbeing status Fourth quintile 771 26.8 406 26.5 Top 20% 1,649 57.2 881 57.6 47 In the World Bank sample, only 90 households with children have a female-head, and 427 households without children have a female-head. In the UNICEF sample, only 52 households with children have a female- head, and 211 households without children have a female-head. 66 Appendix 2: Tables Table 3: Key indicators in UNICEF mobile phone survey Response National Standard Topic Indicator Categories Estimate (%) error Need medical treatment 43.8% 0.06 Difficulties accessing to health care 46.1% 0.10 services Difficulties accessing medicine 18.4% 0.05 Access to Needed medical treatment 40.1% 0.06 health care services Able to access medical treatment 87.9% 0.08 Yes 47.4% 0.06 Willingness to take vaccine No 28.9% 0.06 Not sure 23.7% 0.04 Salary loss / cut 22.5% 0.05 Decrease 65.9% 0.06 Income change due to covid-19 Increase 3.1% 0.02 related reasons Same 31.0% 0.06 Received COVID related support 4.2% 0.01 Cash transfers 42.7% 0.06 Employment Food packs 21.8% 0.06 and income Training and skills Type of support 47.1% 0.06 development In-kind support 40.9% 0.06 Other 24.8% 0.06 Yes 66.6% 0.06 Awareness No 28.2% 0.06 Not sure 5.2% 0.03 In school 6.4% 0.02 At home 72.2% 0.09 At work with me/ 0.0% 0.00 another relative Doing farming or Children living arrangement 0.2% 0.00 household chores (6-14 In house of grand 9.4% 0.07 parents Friend’s house 0.1% 0.00 With caregiver 3.1% 0.03 Other 8.6% 0.07 Television 4.6% 0.01 Child care Radio 22.2% 0.05 Children’s access to learning Smart phone 4.7% 0.01 and playing items (0-5) Computer 52.8% 0.08 Children book 72.2% 0.06 Toys 13.7% 0.04 Television 5.2% 0.01 Radio 39.3% 0.08 Children’s access to learning Smart phone 7.3% 0.02 and playing items (6-14) Computer 66.5% 0.07 Children book 31.7% 0.08 Toys 23.6% 0.06 67 Table 3: Key indicators in UNICEF mobile phone survey / Part 2 Response National Standard Topic Indicator Categories Estimate (%) error Yes 62.7% 0.08 Home learning No 33.8% 0.08 Only some 3.5% 0.02 Received remote learning support 40.1% 0.06 Education Not able to help 13.3% 0.05 Spent more time with Type of remote learning support 69.0% 0.09 children Private tutoring 31.7% 0.09 Other 12.1% 0.08 Concern about learning outcomes 83.0% 0.09 Happy 51.0% 0.06 Sad 16.4% 0.05 Relaxed 8.2% 0.03 Anxious/Worried 30.3% 0.05 Panicking 3.0% 0.02 Life satisfaction for adults Confused 10.0% 0.04 Energized 3.1% 0.01 Tired 17.7% 0.06 Excited about the future 4.1% 0.02 Don't see the point of life 0.3% 0.00 Other 6.6% 0.04 Happy 76.9% 0.06 Sad 12.8% 0.06 Relaxed 14.9% 0.05 Anxious/Worried 9.8% 0.04 Panicking 0.6% 0.00 Life satisfaction for children Confused 5.6% 0.04 Mental Health Energized 13.9% 0.05 Tired 0.8% 0.00 Excited about the future 2.3% 0.01 Don't see the point of life 0.3% 0.00 Other 8.4% 0.05 Changes in child behavior 50.3% 0.06 No new worries expressed 22.0% 0.04 Fear of dying or losing 58.7% 0.06 their relatives/friends Fear of getting COVID-19 or other sickness or of 50.7% 0.06 their loved ones COVID-19 Having less time to spend 14.9% 0.05 related worries/concerns with their friends Online bullying 0.3% 0.00 School closing and not being able to learn, 13.0% 0.03 difficult studying home Other 3.9% 0.02 Not worried 3.3% 0.02 68 Appendix 2: Tables Table 3: Key indicators in UNICEF mobile phone survey / Part 3 Response National Standard Topic Indicator Categories Estimate (%) error No new worries expressed 22.0% 0.04 Fear of dying or losing 58.7% 0.06 their relatives/friends Fear of getting COVID-19 or other sickness or of 50.7% 0.06 their loved ones COVID-19 Having less time to spend 14.9% 0.05 related worries/concerns with their friends Online bullying 0.3% 0.00 School closing and not being able to learn, 13.0% 0.03 difficult studying home Other 3.9% 0.02 Mental Health Not worried 3.3% 0.02 Keep in touch with 32.3% 0.05 relatives and/or friends Focus on the positive side 27.3% 0.05 Stay updated reading news 14.3% 0.05 Spend more time on 1.5% 0.00 Coping strategies for COVID-19 social media/online related worries/concerns Pray 52.8% 0.06 Talking about their worries 16.9% 0.05 Playing/exercising social 28.9% 0.05 and recreational activities Speak to a counsellor 0.2% 0.00 Other 13.0% 0.05 Water and Did not have sufficient Yes, at least once 22.0% 0.05 Sanitation drinking water No, always sufficient 78.0% 0.05 Fixed facility 10.6% 0.03 (sink / tap) in dwelling Fixed facility 40.1% 0.06 (sink / tap) in yard /plot Access to hand washing facilities Mobile object 30.9% 0.05 (bucket / jug / kettle) No handwashing place in 3.4% 0.03 dwelling /yard / plot Handwashing Other 15.0% 0.06 Yes, sufficient soap 73.2% 0.06 and water No, insufficient soap 25.6% 0.06 Sufficient soap and water No, insufficient water 0.8% 0.00 No, insufficient soap 0.4% 0.00 and water 69 Figure 47. Reported emotion of the respondent on data collection day (For households with children only) Excited about future Tired Energized Confused Panicking Anxious/worried Relaxed Sad Happy 0% 20% 40% 60% 80% Female-headed household Male-headed household Excited about future Tired Energized Confused Panicking Anxious/worried Relaxed Sad Happy 0% 10% 20% 30% 40% 50% 60% 70% Rural Other Urban Honiara Source: UNICEF Mobile Phone Survey. 70 Appendix 2: Tables Table 4. Regression analysis for employment outcomes Lost job or working Working for Working for no payment lower pay for higher pay coef/se coef/se coef/se oef/se coef/se coef/se 0.061 0.066 0.036 0.050 -0.004 -0.004 Female respondent (0.059) (0.057) (0.080) (0.075) (0.038) (0.037) -0.043 -0.055 0.015 -0.010 0.040 0.041 Household head (0.071) (0.068) (0.085) (0.080) (0.049) (0.042) Reference: Age 36-45 Age 18-25 -0.013 -0.021 -0.099 -0.141 0.224** 0.243** (0.107) (0.095) (0.111) (0.108) (0.106) (0.100) Age 26-35 0.087 0.094 -0.056 -0.068 0.038 0.045 (0.082) (0.082) (0.093) (0.083) (0.051) (0.046) Age 46-55 -0.132* -0.119 -0.120 -0.098 -0.013 -0.010 (0.075) (0.075) (0.103) (0.091) (0.055) (0.041) Age 56-65 0.214* 0.184* 0.083 0.079 0.025 0.029 (0.120) (0.103) (0.145) (0.114) (0.073) (0.066) Age 66+ 0.240 0.212 -0.175 -0.122 0.021 0.014 (0.163) (0.146) (0.107) (0.083) (0.078) (0.073) Reference: Secondary education No school -0.214** -0.233*** -0.077 -0.139 0.100 0.112 (0.092) (0.089) (0.122) (0.106) (0.098) (0.083) Primary -0.005 -0.001 -0.038 -0.057 -0.038 -0.035 (0.066) (0.061) (0.075) (0.071) (0.058) (0.053) Tertiary -0.018 -0.019 0.019 -0.023 0.034 0.047 (0.062) (0.062) (0.077) (0.070) (0.056) (0.054) Vocational -0.005 -0.048 -0.182 -0.024 -0.027 -0.098 (0.085) (0.115) (0.122) (0.093) (0.073) (0.133) Unknown -0.124 -0.155 -0.284*** -0.123 -0.060 -0.121 (0.100) (0.131) (0.096) (0.111) (0.075) (0.138) Reference: Agriculture Industry -0.065 -0.064 -0.092 -0.056 -0.015 -0.027 (0.084) (0.088) (0.105) (0.100) (0.056) (0.073) Service sector 0.003 -0.005 -0.016 -0.007 0.047 0.040 (0.073) (0.068) (0.081) (0.079) (0.062) (0.070) Full time in January 2020 -0.059 -0.066 0.036 0.000 0.133*** 0.134*** (0.068) (0.066) (0.071) (0.070) (0.047) (0.046) Formal employment -0.263*** -0.248*** -0.090 -0.078 -0.080* -0.084* in January 2020 (0.052) (0.050) (0.070) (0.063) (0.043) (0.043) Reference: Honiara Other urban -0.005 -0.145 0.042 -0.019 -0.067* -0.111* (0.061) (0.122) (0.065) (0.140) (0.039) (0.066) Rural -0.017 -0.143 0.042 -0.053 -0.026 -0.061 (0.051) (0.115) (0.064) (0.133) (0.036) (0.057) Reference: Top 20% Bottom 40% 0.064 0.084 0.001 -0.041 -0.001 0.014 (0.069) (0.075) (0.086) (0.083) (0.063) (0.065) 3rd quintile -0.047 -0.027 0.051 0.025 -0.022 -0.025 (0.066) (0.068) (0.082) (0.081) (0.047) (0.048) 4th quintile -0.053 -0.048 -0.009 -0.030 0.032 0.037 (0.053) (0.052) (0.060) (0.059) (0.039) (0.039) Constant 0.480*** 0.486*** 0.380*** 0.419*** -0.021 -0.022 (0.137) (0.124) (0.140) (0.138) (0.101) (0.106) Province level fixed effects no Yes no yes no yes Observations 4,279 4,279 2,093 2,093 2,093 2,093 Adjusted R2 0.135 0.153 0.034 0.092 0.081 0.096 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 71 Table 5. Non-farm income (over rounds 1 and 2) Round 1 Higher income Lower income No income coef/se coef/se coef/se oef/se coef/se coef/se -0.228 -0.123* -0.104 -0.010 0.022 0.083 Female-headed household (0.149) (0.064) (0.110) (0.101) (0.039) (0.080) Reference: Age 36-45 Age 18-25 0.112 0.160** 0.151 0.124 -0.029 -0.012 (0.076) (0.068) (0.125) (0.101) (0.051) (0.051) Age 26-35 0.180 0.164 0.071 0.044 -0.077 -0.069 (0.131) (0.109) (0.103) (0.103) (0.065) (0.059) Age 46-55 0.003 0.000 0.461*** 0.439*** -0.021 -0.017 (0.039) (0.036) (0.133) (0.129) (0.068) (0.060) Age 56-65 -0.081 -0.045 -0.056 -0.133 -0.043 -0.053 (0.062) (0.043) (0.098) (0.099) (0.062) (0.070) Age 66+ -0.001 -0.058 -0.179 -0.172 -0.056 -0.046 (0.038) (0.069) (0.115) (0.122) (0.047) (0.052) Reference: Honiara Other urban -0.026 -0.173** -0.079 -0.212 -0.023 0.041 (0.039) (0.083) (0.097) (0.146) (0.033) (0.069) Rural 0.029 -0.095 -0.166** -0.324** 0.007 0.057 (0.050) (0.065) (0.082) (0.128) (0.048) (0.073) Reference: Top 20% Bottom 40% 0.034 0.007 -0.209* -0.137 -0.025 -0.002 (0.080) (0.103) (0.109) (0.121) (0.045) (0.053) 3rd quintile -0.086* -0.107** -0.054 -0.069 -0.046 -0.040 (0.051) (0.047) (0.116) (0.104) (0.043) (0.046) 4th quintile -0.050 -0.061 0.060 -0.002 0.018 0.028 (0.041) (0.042) (0.090) (0.081) (0.049) (0.050) Constant 0.055* 0.043 0.486*** 0.508*** 0.088** 0.072** (0.032) (0.034) (0.074) (0.072) (0.043) (0.036) Includes provinces level fixed no yes no yes no yes effects Observations 2,664 2,664 2,664 2,664 2,664 2,664 Adjusted R2 0.159 0.221 0.255 0.324 0.036 0.095 72 Appendix 2: Tables Table 5. Non-farm income (over rounds 1 and 2) / Part 2 Round 2 Higher income Lower income No income coef/se coef/se coef/se oef/se coef/se coef/se 0.046 0.029 0.099 0.084 -0.065 -0.039 (0.062) (0.057) (0.090) (0.084) (0.043) (0.031) 0.145* 0.139** 0.417* 0.431** 0.012 -0.002 (0.086) (0.069) (0.214) (0.211) (0.019) (0.010) -0.019 -0.042 0.459*** 0.490*** 0.050* 0.030 (0.062) (0.061) (0.155) (0.160) (0.027) (0.021) -0.023 -0.026 0.441*** 0.516*** 0.030 -0.063 (0.054) (0.056) (0.151) (0.164) (0.021) (0.048) -0.077 -0.072 0.335* 0.348* 0.127 0.108 (0.068) (0.066) (0.194) (0.201) (0.085) (0.080) -0.097* -0.112** 0.258 0.384* 0.356* 0.202* (0.059) (0.056) (0.210) (0.202) (0.193) (0.104) -0.008 -0.076 0.038 0.248 -0.021 -0.142** (0.057) (0.084) (0.084) (0.273) (0.029) (0.072) -0.044 -0.085 -0.089 0.115 0.092 -0.053 (0.125) (0.076) (0.145) (0.278) (0.065) (0.065) -0.032 0.101 0.289 0.193 -0.076 -0.056 (0.141) (0.160) (0.193) (0.186) (0.051) (0.037) 0.074 0.094 0.233* 0.205 -0.066* -0.058** (0.082) (0.080) (0.126) (0.130) (0.036) (0.030) 0.042 0.057 0.094 0.096 0.038 0.022 (0.059) (0.056) (0.088) (0.084) (0.056) (0.040) 0.082 0.084 0.050 0.030 -0.007 0.023 (0.066) (0.061) (0.157) (0.158) (0.043) (0.029) no yes no yes no yes 2,882 2,882 2,882 2,882 2,882 2,882 0.076 0.142 0.148 0.186 0.277 0.458 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Rounds one and two of the World Bank High Frequency Mobile Phone Survey. 73 Table 6. Main activity for agricultural household Food crops Food crops for sale Cash crops for home production in local market for sale or export coef/se coef/se coef/se oef/se coef/se coef/se Female-headed 0.497*** 0.409*** -0.295** -0.214*** 0.014 0.008 household (0.116) (0.087) (0.138) (0.071) (0.023) (0.028) Reference: Age 36-45 Age 18-25 -0.004 -0.078 -0.209 -0.165* 0.047 0.047 (0.128) (0.123) (0.137) (0.095) (0.056) (0.063) Age 26-35 -0.058 -0.130 -0.181 -0.116 0.016 0.015 (0.114) (0.106) (0.140) (0.098) (0.021) (0.027) Age 46-55 -0.088 -0.231* -0.308 -0.228** -0.005 -0.019 (0.175) (0.136) (0.194) (0.116) (0.010) (0.020) Age 56-65 0.078 -0.152 -0.164 -0.005 -0.010 -0.005 (0.140) (0.128) (0.150) (0.096) (0.013) (0.028) Age 66+ 0.393** 0.153 -0.220 -0.110 -0.005 -0.022 (0.191) (0.160) (0.202) (0.190) (0.024) (0.040) Reference: Honiara Other urban -0.057 -0.208 0.104* -0.074 0.019 0.146 (0.079) (0.246) (0.058) (0.100) (0.022) (0.129) Rural -0.237** -0.443* 0.052 -0.059 0.022 0.149 (0.099) (0.237) (0.066) (0.093) (0.016) (0.138) Reference: Top 20% Bottom 40% 0.081 0.057 0.100 0.145 -0.017 -0.018 (0.131) (0.134) (0.129) (0.127) (0.033) (0.031) 3rd quintile 0.356*** 0.358*** -0.076 -0.076 -0.025 -0.028 (0.106) (0.092) (0.064) (0.061) (0.018) (0.018) 4th quintile 0.186** 0.173** -0.044 -0.044 -0.011 -0.016 (0.072) (0.074) (0.053) (0.053) (0.021) (0.022) Constant 0.467*** 0.586*** 0.365*** 0.277*** -0.001 0.004 (0.116) (0.100) (0.141) (0.085) (0.033) (0.043) Includes provinces level fixed no yes no Yes no yes effects Observations 937 937 937 937 937 937 Adjusted R2 0.282 0.325 0.196 0.270 0.014 0.028 74 Appendix 2: Tables Table 6. Main activity for agricultural household / Part 2 Livestock for home Livestock Fish for home Fish consumption for sale consumption for sale coef/se coef/se coef/se oef/se coef/se coef/se coef/se coef/se 0.001 0.002 -0.037 -0.044 -0.007* -0.008 -0.175* -0.157** (0.015) (0.014) (0.029) (0.034) (0.003) (0.006) (0.099) (0.062) 0.006 0.004 0.115** 0.114** -0.003 -0.004 0.042 0.073* (0.019) (0.017) (0.055) (0.053) (0.005) (0.005) (0.043) (0.037) -0.005 -0.007 0.123** 0.111** 0.014 0.012 0.081 0.101 (0.012) (0.010) (0.058) (0.052) (0.012) (0.013) (0.058) (0.063) -0.008 -0.013 -0.010 -0.014 0.005 0.003 0.412* 0.496*** (0.014) (0.016) (0.026) (0.039) (0.006) (0.009) (0.228) (0.161) -0.009 -0.005 0.010 0.008 0.002 -0.001 0.081 0.127** (0.010) (0.008) (0.028) (0.061) (0.004) (0.007) (0.055) (0.054) -0.022* -0.021 -0.096*** -0.078** -0.006 -0.011 -0.026 0.102 (0.013) (0.016) (0.035) (0.033) (0.006) (0.011) (0.045) (0.091) -0.005 -0.009 -0.044 0.249 0.022** 0.011 -0.025 -0.088** (0.010) (0.013) (0.039) (0.231) (0.010) (0.013) (0.029) (0.044) 0.013 0.006 0.087* 0.348 0.019* 0.008 0.057 0.015 (0.010) (0.011) (0.045) (0.235) (0.011) (0.014) (0.040) (0.048) -0.029* -0.023 -0.127** -0.106* -0.014 -0.015 0.046 -0.001 (0.015) (0.014) (0.057) (0.058) (0.011) (0.014) (0.053) (0.049) -0.019 -0.020 -0.132*** -0.135*** -0.001 -0.001 -0.056 -0.051 (0.015) (0.015) (0.046) (0.045) (0.008) (0.009) (0.034) (0.034) -0.016 -0.018 -0.047 -0.046 -0.006 -0.006 -0.036 -0.016 (0.010) (0.011) (0.043) (0.044) (0.009) (0.009) (0.035) (0.034) 0.025 0.027 0.088** 0.095** -0.002 -0.000 0.007 -0.034 (0.018) (0.018) (0.042) (0.041) (0.006) (0.007) (0.050) (0.041) no yes no yes no yes no yes 937 937 937 937 937 937 937 937 -0.003 0.053 0.066 0.087 0.005 -0.003 0.370 0.491 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Rounds One and Round Two of the World Bank High Frequency Mobile Phone Survey. 75 Table 7. Expected agricultural income Higher Lower No expected income expected income expected income coef/se coef/se coef/se oef/se coef/se coef/se Female-headed -0.061 -0.044 0.049 0.069 -0.026 -0.012 household (0.050) (0.046) (0.086) (0.094) (0.039) (0.048) Reference: Age 36-45 Age 18-25 0.104** 0.126** -0.497*** -0.447*** 0.093* 0.097* (0.045) (0.064) (0.131) (0.104) (0.056) (0.053) Age 26-35 0.175* 0.201* -0.529*** -0.492*** 0.026 0.020 (0.090) (0.109) (0.130) (0.102) (0.037) (0.042) Age 46-55 0.038 0.090 -0.712*** -0.738*** 0.004 0.003 (0.035) (0.062) (0.164) (0.130) (0.032) (0.090) Age 56-65 0.046 0.099 -0.797*** -0.557*** 0.035 0.200* (0.044) (0.096) (0.144) (0.109) (0.049) (0.110) Age 66+ -0.029 0.021 -0.102 -0.069 0.227 0.230 (0.070) (0.086) (0.268) (0.252) (0.229) (0.247) Reference: Honiara Other urban -0.103* -0.147* 0.104* -0.074 0.092 -0.124 (0.056) (0.086) (0.058) (0.100) (0.093) (0.133) Rural 0.009 -0.023 0.052 -0.059 -0.108 -0.316** (0.070) (0.096) (0.066) (0.093) (0.102) (0.142) Reference: Top 20% Bottom 40% -0.065 -0.056 0.062 0.134 -0.020 -0.013 (0.066) (0.070) (0.126) (0.121) (0.105) (0.105) 3rd quintile -0.040 -0.025 0.089 0.091 -0.049 -0.056 (0.064) (0.071) (0.101) (0.104) (0.097) (0.099) 4th quintile 0.006 0.028 0.110 0.140 -0.094 -0.089 (0.062) (0.070) (0.088) (0.091) (0.092) (0.092) Constant 0.074 0.038 0.651*** 0.589*** 0.150*** 0.132** (0.054) (0.077) (0.123) (0.095) (0.055) (0.053) Includes provinces level no yes No yes no yes fixed effects Observations 634 634 634 634 634 634 Adjusted R2 0.069 0.068 0.346 0.388 0.103 0.108 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 76 Appendix 2: Tables Table 8. Financial Anxiety and Economic Outlook (Ordered Logit regressions) Financial Anxiety Economic Outlook coef/se coef/se coef/se oef/se coef/se coef/se Female respondent 1.237*** 1.089** 0.882* -0.628 -0.418 -0.0342 (0.474) (0.452) (0.522) (0.522) (0.542) (0.327) Respondent is 0.792 0.867* 0.618 0.150 0.433 0.546* household head (0.510) (0.461) (0.565) (0.482) (0.469) (0.295) Respondent is -1.185*** -1.004** -0.384 1.784*** 1.713*** -0.699 household head (0.457) (0.424) (0.483) (0.612) (0.611) (0.479) Reference: Age 36-45 Age 18-25 0.255 0.721* -0.607 -0.997** -1.238*** -0.160 (0.530) (0.415) (0.523) (0.445) (0.420) (0.491) Age 26-35 -0.423 -0.114 -0.200 -0.366 -0.624 0.272 (0.689) (0.549) (0.429) (0.440) (0.410) (0.409) Age 46-55 0.295 0.250 0.252 -1.327** -2.315*** -0.168 (0.828) (0.527) (0.512) (0.532) (0.624) (0.492) Age 56-65 0.904 1.566* -0.254 -0.472 -0.549 0.835 (1.095) (0.913) (0.862) (0.907) (0.563) (0.673) Age 66+ 0.337 1.096 6.447*** -0.280 0.408 2.278* (0.741) (1.030) (2.149) (0.631) (0.809) (1.195) Reference: Secondary education No school 2.465*** 2.936*** 2.513*** 0.823 0.660 -1.643** (0.685) (0.631) (0.773) (0.718) (0.656) (0.733) Primary -0.00798 -0.126 -0.817 -0.0879 0.0545 0.141 (0.378) (0.338) (0.564) (0.293) (0.302) (0.331) Tertiary 0.318 0.366 -0.0726 0.0144 -0.0830 -0.174 (0.217) (0.255) (0.355) (0.241) (0.277) (0.351) Vocational 0.231 -1.165 0.297 2.407** 1.665* 4.269*** (0.689) (0.734) (0.701) (1.052) (0.953) (1.309) Reference: Agriculture Industry 0.662 0.180 (0.537) (0.607) -0.425 0.419 Service sector (0.453) (0.384) 0.965* 0.560 Full time in January 2020 (0.556) (0.343) Formal employment -0.152 0.326 in January 2020 (0.372) (0.341) Reference: Honiara Other urban 0.00668 -0.534 0.303 -0.321 -0.951* -0.238 (0.242) (0.681) (0.411) (0.299) (0.530) (0.495) Rural 0.0712 0.236 0.980** -0.205 -1.320** -0.146 (0.508) (0.757) (0.434) (0.532) (0.594) (0.437) Reference: Top 20% Bottom 40% 0.0589 0.725 -1.391** 0.243 0.148 0.0617 (0.568) (0.627) (0.571) (0.511) (0.483) (0.530) 3rd quintile 0.352 0.782 -0.0881 0.517 0.469 -0.147 (0.474) (0.540) (0.562) (0.557) (0.532) (0.681) 4th quintile -0.103 -0.102 -0.534 -0.226 -0.349 -0.283 (0.263) (0.299) (0.435) (0.283) (0.286) (0.438) Cuts Cut 1 0.477 0.809 0.475 -5.373*** -5.596*** -4.566*** (0.643) (0.503) (0.910) (0.629) (0.647) (1.159) Cut 2 2.786*** 3.461*** 3.233*** -2.846*** -3.023*** -2.326** (0.778) (0.594) (1.038) (0.605) (0.625) (1.077) Cut 3 4.405*** 5.282*** 6.203*** -1.112** -1.202** 0.0177 (0.810) (0.696) (1.802) (0.518) (0.481) (0.775) Cut 4 0.914* 1.086** 2.451*** (0.517) (0.489) (0.831) Province level fixed effects no yes no no yes no Observations 2,877 2,877 1,135 2,877 2,877 1,135 note: *** p<0.01, ** p<0.05, * p<0.1 77 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. Table 9. Financial services Expressed the need Expressed the need to go to a bank to go to a money agent coef/se coef/se coef/se oef/se -0.028 -0.024 -0.024 -0.020 Female-headed household (0.035) (0.041) (0.030) (0.035) Reference: Age 36-45 Age 18-25 0.047 0.048 0.017 0.010 (0.041) (0.048) (0.021) (0.027) Age 26-35 -0.018 0.020 0.050* 0.089** (0.032) (0.040) (0.030) (0.042) Age 46-55 0.092 0.176 0.127 0.186 (0.119) (0.139) (0.117) (0.140) Age 56-65 -0.046 -0.050 -0.025 -0.032 (0.040) (0.047) (0.027) (0.031) Age 66+ -0.061 -0.045 0.022 0.038 (0.067) (0.078) (0.057) (0.071) Reference: Honiara Other urban 0.029 -0.032 0.019 0.032 (0.045) (0.066) (0.059) (0.063) Rural -0.146** -0.155* -0.201* -0.161 (0.069) (0.084) (0.110) (0.104) Reference: Top 20% Bottom 40% -0.487*** -0.514*** -0.057 -0.073 (0.073) (0.069) (0.100) (0.092) 3rd quintile -0.451*** -0.447*** -0.045 -0.037 (0.066) (0.066) (0.114) (0.110) 4th quintile -0.150*** -0.155*** 0.036 0.026 (0.047) (0.045) (0.055) (0.052) Constant 0.697*** 0.678*** 0.276*** 0.260*** (0.036) (0.040) (0.030) (0.032) Includes provinces level fixed effects no yes no yes Observations 2,882 2,882 2,882 2,882 Adjusted R2 0.342 0.360 0.136 0.157 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 78 Appendix 2: Tables Table 10. Ability to access needed care Urgent care Routine care Preventive care coef/se coef/se coef/se 0.012 -0.105 0.086*** Female-headed household (0.027) (0.094) (0.029) Share of household members under age 2 -0.043* -0.000 -0.040* (0.024) (0.030) (0.022) -0.033* -0.091 0.007 Share of household members age 3-5 (0.020) (0.069) (0.015) 0.013 0.019 0.004 Share of household members age 6-14 (0.009) (0.025) (0.011) -0.055 -0.015 -0.080** Share of household members over age 65 (0.035) (0.053) (0.036) 0.003 0.005 0.004 Household size (0.008) (0.022) (0.009) Reference: Top 20% Bottom 40% 0.019 0.019 -0.551*** (0.049) (0.049) (0.159) 3rd quintile 0.088* 0.088* -0.037 (0.046) (0.046) (0.068) 4th quintile 0.070** 0.070** -0.027 (0.035) (0.035) (0.026) Rural rewsience 0.021 0.021 0.009 (0.061) (0.061) (0.079) Reference: Honiara Choiseul -0.433 0.181* 0.056 (0.315) (0.102) (0.054) Westwern -0.014 0.150** -0.056 (0.063) (0.073) (0.044) Isabel -0.017 0.275** 0.125** (0.080) (0.134) (0.064) Central -0.405* -0.478 -0.084 (0.242) (0.357) (0.192) Rennell-Bellona -0.867*** -0.672*** 0.173* (0.097) (0.237) (0.089) Guadalcanal -0.082 0.097 -0.016 (0.093) (0.205) (0.099) Malaita 0.002 0.119 0.514*** (0.065) (0.161) (0.162) Makira-Ulawa -0.253 0.096 0.006 (0.292) (0.200) (0.075) Temotu -0.017 0.362** -0.061 (0.075) (0.175) (0.050) Constant 0.943*** 0.922*** 0.972*** (0.061) (0.106) (0.041) N 381 201 290 Adjusted R2 0.443 0.445 0.476 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 79 Table 11. Change in community safety situation in previous six months [urban only] BETTER Community Theft Damage to Alcohol and Intimidation Land Domestic Trust property drug abuse by police disputes abuse coef/se coef/se coef/se oef/se coef/se coef/se coef/se 0.038 0.111** 0.053 -0.002 -0.046 0.105* 0.052 Female respondent (0.047) (0.053) (0.049) (0.041) (0.048) (0.056) (0.051) Reference: Age 36-45 Age 18-25 -0.015 -0.109* -0.066 -0.002 0.000 0.154** 0.023 (0.056) (0.061) (0.062) (0.053) (0.060) (0.066) (0.059) Age 26-35 -0.055 0.014 0.025 -0.074 -0.060 0.028 -0.096* (0.060) (0.059) (0.063) (0.053) (0.059) (0.068) (0.055) Age 46-55 -0.052 -0.082 -0.103 -0.015 -0.049 -0.027 -0.027 (0.056) (0.075) (0.073) (0.057) (0.067) (0.066) (0.073) Age 56-65 -0.035 0.050 0.205*** 0.163* 0.135 0.060 0.153 (0.080) (0.086) (0.074) (0.092) (0.097) (0.099) (0.099) Age 66+ -0.058 -0.086 -0.159 -0.231*** -0.056 0.044 -0.010 (0.140) (0.147) (0.130) (0.089) (0.136) (0.155) (0.143) Reference: Secondary education No school -0.265*** -0.042 0.329** 0.056 0.346** 0.229 0.188 (0.102) (0.152) (0.141) (0.146) (0.141) (0.198) (0.148) Primary -0.003 0.020 0.044 -0.052 -0.020 0.200** -0.111 (0.076) (0.073) (0.073) (0.065) (0.074) (0.085) (0.071) Tertiary 0.037 -0.011 0.023 0.063 0.082 0.020 0.005 (0.056) (0.060) (0.061) (0.052) (0.061) (0.066) (0.062) Vocational -0.049 -0.104 -0.284*** -0.092 -0.053 0.050 0.131 (0.109) (0.119) (0.098) (0.090) (0.115) (0.155) (0.113) Reference: Honiara Other urban -0.105** -0.105* -0.050 -0.041 -0.009 -0.207*** -0.083 (0.053) (0.058) (0.059) (0.056) (0.059) (0.061) (0.056) Reference: Top 20% Bottom 40% -0.284*** -0.172 -0.263 -0.264*** -0.371*** -0.593*** -0.226 (0.064) (0.184) (0.183) (0.060) (0.085) (0.102) (0.162) 3rd quintile 0.320** 0.192* 0.101 -0.037 -0.147 0.002 -0.234*** (0.137) (0.105) (0.120) (0.102) (0.115) (0.135) (0.086) 4th quintile -0.009 0.010 -0.039 -0.025 -0.015 -0.031 -0.051 (0.056) (0.062) (0.059) (0.049) (0.060) (0.069) (0.062) Constant 0.344*** 0.434*** 0.492*** 0.349*** 0.492*** 0.368*** 0.488*** (0.059) (0.061) (0.060) (0.056) (0.058) (0.065) (0.056) Observations 1,981 1,939 1,917 1,968 1,919 1,797 1,908 Adjusted R2 0.088 0.071 0.083 0.055 0.071 0.113 0.091 80 Appendix 2: Tables Table 11. Change in community safety situation in previous six months [urban only] / Part 2. WORSE Community Theft Damage to Alcohol and Intimidation Land Domestic Trust property drug abuse by police disputes abuse coef/se coef/se coef/se oef/se coef/se coef/se coef/se 0.224 0.495*** -0.040 0.071 0.063** -0.022 -0.045 (0.167) (0.190) (0.030) (0.051) (0.031) (0.039) (0.034) -0.117*** 0.057 0.027 -0.017 -0.020 -0.040 -0.037 (0.039) (0.057) (0.048) (0.057) (0.044) (0.048) (0.047) -0.065* -0.017 -0.048 0.048 -0.045 0.030 -0.017 (0.038) (0.050) (0.041) (0.060) (0.038) (0.047) (0.044) -0.080* 0.028 -0.074* -0.076 0.033 0.066 -0.035 (0.045) (0.089) (0.041) (0.071) (0.077) (0.080) (0.057) -0.016 0.017 -0.114*** -0.176* 0.051 0.028 -0.089 (0.068) (0.082) (0.043) (0.091) (0.072) (0.082) (0.069) 0.111 -0.104 -0.097 -0.097 -0.120*** -0.012 -0.139* (0.144) (0.096) (0.069) (0.141) (0.035) (0.085) (0.072) -0.160** -0.112 -0.143*** -0.210 0.022 -0.141** -0.005 (0.080) (0.088) (0.044) (0.129) (0.083) (0.070) (0.088) -0.071* -0.068 -0.071* -0.022 -0.110*** -0.115*** -0.034 (0.042) (0.055) (0.039) (0.076) (0.039) (0.044) (0.040) -0.063 0.023 -0.023 -0.022 -0.022 0.083 0.009 (0.046) (0.059) (0.043) (0.059) (0.043) (0.063) (0.050) -0.154*** -0.019 0.004 0.071 -0.001 -0.161*** 0.017 (0.047) (0.110) (0.076) (0.145) (0.075) (0.040) (0.082) 0.007 0.049 0.002 0.004 0.098** 0.222*** 0.086* (0.038) (0.052) (0.038) (0.061) (0.044) (0.052) (0.045) 0.001 0.004 -0.091 -0.090 -0.029 0.010 -0.094 (0.061) (0.227) (0.065) (0.175) (0.061) (0.076) (0.069) -0.090** -0.079 0.028 0.248** 0.002 -0.006 -0.064 (0.042) (0.079) (0.070) (0.125) (0.058) (0.060) (0.060) 0.000 -0.045 -0.028 -0.010 -0.009 0.017 -0.049 (0.049) (0.050) (0.037) (0.059) (0.039) (0.052) (0.039) 0.205*** 0.272*** 0.226*** 0.363*** 0.109*** 0.143*** 0.240*** (0.039) (0.053) (0.041) (0.057) (0.042) (0.047) (0.042) 1,981 1,939 1,917 1,968 1,919 1,797 1,908 0.053 0.028 0.033 0.078 0.069 0.105 0.028 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 81 Table 12. Change in community safety situation in previous six months [rural]. BETTER Community Theft Damage to Alcohol and Intimidation Land Domestic Trust property drug abuse by police disputes abuse coef/se coef/se coef/se oef/se coef/se coef/se coef/se -0.363** -0.608*** -0.348* -0.514*** -0.778*** -0.639*** -0.246 Female respondent (0.185) (0.153) (0.189) (0.191) (0.102) (0.144) (0.197) Reference: Age 36-45 Age 18-25 -0.353*** 0.397** 0.130 -0.182 -0.116 0.395* 0.444* (0.136) (0.198) (0.179) (0.238) (0.120) (0.204) (0.238) Age 26-35 -0.352* 0.086 -0.346* -0.231 -0.050 0.070 0.379 (0.182) (0.150) (0.182) (0.226) (0.109) (0.140) (0.234) Age 46-55 -0.459** -0.276 -0.023 -0.524*** -0.466*** 0.079 -0.119 (0.206) (0.219) (0.207) (0.191) (0.155) (0.176) (0.246) Age 56-65 -0.217 0.035 -0.277 -0.638*** -0.445*** -0.186 0.131 (0.273) (0.270) (0.225) (0.205) (0.172) (0.255) (0.320) Age 66+ -0.667** -0.131 -0.490** -0.406* -0.441*** -0.262 0.137 (0.162) (0.205) (0.220) (0.209) (0.120) (0.179) (0.246) Reference: Secondary education No school -0.204 -0.496** 0.054 -0.169 -0.421*** -0.052 -0.052 (0.158) (0.207) (0.130) (0.170) (0.158) (0.169) (0.169) Primary -0.180 -0.201 -0.398*** 0.180 -0.118 -0.233 -0.233 (0.119) (0.160) (0.105) (0.163) (0.093) (0.166) (0.166) Tertiary -0.196 -0.362** -0.152 0.009 -0.328*** 0.006 0.006 (0.180) (0.156) (0.182) (0.171) (0.108) (0.161) (0.161) Vocational -0.507*** -0.137 -0.643*** -0.163 0.239 -0.155 -0.155 (0.149) (0.189) (0.147) (0.193) (0.258) (0.193) (0.193) Reference: Top 20% Bottom 40% -0.069 0.150 0.066 0.020 -0.147 0.344*** 0.370*** (0.084) (0.135) (0.118) (0.122) (0.102) (0.118) (0.110) 3rd quintile -0.154 -0.537*** -0.164 -0.467*** -0.705*** -0.383*** -0.379** (0.107) (0.129) (0.120) (0.164) (0.146) (0.112) (0.150) 4th quintile 0.223** 0.181 0.214* 0.131 -0.163* -0.117 0.188* (0.093) (0.123) (0.111) (0.119) (0.094) (0.119) (0.103) Constant 0.779*** 0.771*** 0.877*** 0.747*** 1.115*** 0.632*** 0.329 (0.203) (0.213) (0.197) (0.186) (0.123) (0.216) (0.245) Observations 976 960 964 968 943 947 946 Adjusted R2 0.335 0.478 0.530 0.387 0.577 0.551 0.447 82 Appendix 2: Tables Table 12. Change in community safety situation in previous six months [rural] / Part 2. WORSE Community Theft Damage to Alcohol and Intimidation Land Domestic Trust property drug abuse by police disputes abuse coef/se coef/se coef/se oef/se coef/se coef/se coef/se 0.224 0.495*** 0.216 0.583*** 0.105 0.302 0.349** (0.167) (0.190) (0.171) (0.183) (0.066) (0.201) (0.174) -0.320* -0.239 -0.343* 0.291 0.403*** -0.639*** -0.352* (0.188) (0.206) (0.204) (0.231) (0.152) (0.164) (0.186) -0.274 0.213 -0.273 0.297 0.063 -0.391* -0.455** (0.199) (0.222) (0.204) (0.217) (0.058) (0.221) (0.201) -0.146 0.551** -0.306* 0.491** -0.001 -0.299 -0.285 (0.172) (0.242) (0.175) (0.201) (0.082) (0.219) (0.176) -0.186 0.248 0.407* 0.700*** 0.387* -0.021 0.117 (0.173) (0.277) (0.239) (0.204) (0.209) (0.307) (0.274) -0.233* 0.442** -0.255* 0.705*** -0.026 -0.782*** -0.133 (0.122) (0.206) (0.148) (0.170) (0.098) (0.257) (0.191) -0.052 0.447** 0.133 0.466** -0.163 -0.084 0.178* (0.169) (0.223) (0.131) (0.185) (0.111) (0.155) (0.102) -0.233 0.211** 0.018 0.049 -0.248** 0.098 0.241** (0.166) (0.102) (0.107) (0.133) (0.098) (0.137) (0.105) 0.006 -0.040 -0.043 0.134 -0.043 -0.099 0.143 (0.161) (0.119) (0.076) (0.148) (0.068) (0.181) (0.153) -0.155 0.326 -0.130 -0.120 0.034 -0.133 -0.015 (0.193) (0.212) (0.151) (0.218) (0.151) (0.219) (0.194) -0.090 -0.045 -0.048 0.163 0.198** -0.336*** -0.084 (0.067) (0.115) (0.099) (0.130) (0.098) (0.125) (0.105) 0.020 0.544*** 0.075 0.592*** 0.098 0.323* 0.066 (0.088) (0.139) (0.119) (0.183) (0.083) (0.180) (0.112) -0.078 -0.114 -0.010 -0.090 -0.036 -0.067 -0.024 (0.061) (0.111) (0.096) (0.113) (0.063) (0.163) (0.106) 0.223** -0.134 0.239* -0.228* 0.001 0.670** 0.180 (0.111) (0.148) (0.142) (0.135) (0.105) (0.262) (0.197) 976 960 964 968 943 947 946 0.272 0.438 0.271 0.390 0.554 0.451 0.479 note: *** p<0.01, ** p<0.05, * p<0.1 Source: Round Two of the World Bank High Frequency Mobile Phone Survey. 83 84 85 86