WELFARE IMPACTS OF THE COVID-19 PANDEMIC IN MOLDOVA THROUGH THE LABOR AND MIGRATION CHANNELS Abstract This paper describes the contribution of migration and remittances to household welfare in Moldova based on recent data, as well as the implications of the COVID-19 pandemic on household welfare working through the channels of labor market shocks, return migration and remittances – a salient characteristic of the Moldovan economy. Simulations suggest that in the absence of any direct government intervention to households, return migrants, and workers, the poverty rate could increase by as much as 2.9 percentage points on account of the COVID-19 pandemic, by pushing close to an additional 100,000 people into poverty, primarily through employment losses in pandemic affected sectors. Analysis of the government’s social assistance package to households, return migrants and workers suggests that under certain assumptions on duration, coverage and generosity, the effects of the COVID-19 crisis can be partially, but not fully, mitigated. The impacts of the pandemic associated with return migration and pandemic impacts on remittances are estimated to be more muted, relative to labor market shocks, as initial expectations regarding the magnitude of return migration and remittances decline do not appear to have materialized. Alexandru Cojocaru, Kristina Vaughan and Ganesh Kumar Seshan May 2021 Contents Executive summary ....................................................................................................................................... 2 I. Introduction ............................................................................................................................................... 5 II. Migration and remittances as a key livelihood source in Moldova .......................................................... 6 III. Key channels of welfare impacts of COVID-19 pandemic ........................................................................ 8 The labor market channel ......................................................................................................................... 9 The migration / remittances channel ..................................................................................................... 11 IV. Data description and empirical strategy................................................................................................ 13 Data ......................................................................................................................................................... 13 Simulations setup.................................................................................................................................... 14 V. Main results: unmitigated effects of the pandemic on poverty and the profile of the new poor ......... 19 Estimated impacts of the pandemic on the incidence and depth of poverty ........................................ 19 Profile of the new vs existing poor ......................................................................................................... 21 VI. The welfare effects of mitigating policy measures ................................................................................ 23 VII. How long will the return migrants stay and the welfare effects of their re-migrating? ...................... 27 VIII. Experience of return migrants ............................................................................................................. 30 IX. Conclusions ............................................................................................................................................ 34 References .................................................................................................................................................. 37 Appendix: Additional figures and tables ..................................................................................................... 39 1 Executive summary As a result of the COVID-19 pandemic, the poverty rate in Moldova is projected to increase from 12.9 percent to 15.5 percent using the upper middle-income poverty line of US$5.50 PPP or MDL 43.80 PPP per day pushing a potential 93,000 people into poverty. The increase is based on the simulated effect of COVID-19 on the labor market, remittances and return migration under certain assumptions and in the absence of any direct COVID-19 government intervention to households, return migrants and workers. Projected increases in poverty are larger in rural areas (3.0 percentage points) than in urban areas (2.2 percentage points), primarily due to the increased impact of return migration in rural areas (Figure 1). COVID-19’s adverse effects on employment account for the largest share of the simulated increase in poverty in both urban and rural areas of 1.8 and 1.5 percentage point increases, respectively. This is due, in part, to the fact that the initial expectations related to the magnitude of return migration and the drop in the flow of remittances on account of pandemic mitigation measures did not fully materialize. Fig 1: Simulated effects on the national and urban/rural poverty headcount across various COVID-19 scenarios and the baseline scenario 25 21.1 19.6 20 18.1 18.3 18.3 Poverty headcount rate 15.5 14.4 15 12.9 13.3 13.1 10 8.3 7.8 6.9 6.3 6.1 5 0 Baseline All All+SA(2) All+SA(4) All+SA(6) National Urban Rural Notes: Bars shows estimated poverty rates. “Baseline” is the poverty rate in Moldova (based on a poverty line of US$5.5 per person per day) pre-COVID. “All” is the simulated poverty rate with all the considered adverse effects of COVID included but without government response measures. “All+SA(2)” is the scenario in which the GoM provides two months of income support. “All+SA(4)” is the scenario in which the GoM provides four months of income support. “All+SA(6)” is the scenario in which the GoM provides six months of income support. The increase in the poverty rate can be partially mitigated by implementing social support measures. Incorporating approximations to the government’s social support package into the simulations suggests that with sufficient duration, coverage and generosity of unemployment and Guaranteed Minimum Income benefits, increases over baseline poverty can be tempered. It is important to note, however, that these are projections and may not accurately reflect the actual implemented benefits and the result should be interpreted in that context. 2 In addition to increases in the poverty rate, those who were already impoverished before COVID-19 risk falling deeper into poverty as a result of the crisis. The poverty gap, a measure of the depth of poverty, is simulated to increase from 2.3 to 3.5 due to COVID-19’s combined effects on the labor market, remittances and return migration and in the absence of any direct government intervention to households, return migrants, and workers. When accounting for the government’s proposed social policy package to households, return migrants and workers, with certain assumptions about duration, generosity and coverage, the magnitude of the increase can be tempered but not reversed. Movements in the urban/rural poverty gap largely mirrored those in the urban/rural poverty headcount rate with the magnitude of increase in the rural poverty gap outpacing that of the urban poverty gap, with both being largely driven by the labor market effects of COVID-19. The “new poor”, people who are projected to be pushed into poverty as a result of COVID-19, have different characteristics than the “traditionally poor” (those were below the poverty line even pre- COVID). The “new poor” tend to be of working-age, tend to have fewer children and were typically employed in the services sector before the pandemic struck, have higher levels of educational attainment and were more likely to be employees (at least before the pandemic) rather than self-employed. A fifth of the new poor are concentrated in the capital city Chisinau region and a third of the working new poor were concentrated in the agriculture sector. The profile of the ‘new poor’ stands in stark contrast to the “traditionally” or existing poor that had more children, are largely working in agriculture, have lower educational attainment and tend to be self-employed or are elderly. Less than 3 percent of the existing poor were concentrated in the Chisinau region. Adequate social protection measures could be implemented to prevent the existing poor from falling deeper into poverty and to prevent the emergence of the “new poor”. The existing poor and a third of the new poor are disproportionately self-employed individuals in the agriculture sector. In addition to the COVID-19 shock, individuals in the agriculture sector also experienced the shock of a severe drought in the first half of 2020 that affected agriculture productivity and employment. Given their employment dynamics, they may have limited means to self-insure against shocks and may be inadequately covered by existing social support systems, putting them at increased risk of falling deeper into poverty. Effective temporary unemployment insurance could help protect the incomes of those who lost their wage-job because of the pandemic, particularly since their characteristics are likely to exclude them from traditional social protection programs As economies around Europe recover, it is expected that return migrants will remigrate back to their former host countries and with the resumption of remittances, help alleviate some of the adverse impacts of the crisis on households. Based on surveyed intentions to return, a 67 percent re-migration rate tempers the impact of migration on poverty from an increase of 0.7 to an increase of 0.2 percentage points over baseline poverty. Consequently, under the “All” scenario, the increase is tempered from a 2.7 to a 2.1 percentage point increase over baseline poverty. Further recovery in the labor market and more broadly in remittances is likely to lead to a further mitigation of the effects of COVID-19 on poverty. However, some returnees may not be able to re-migrate in the near future and likely need additional government support in the interim, both in terms of job search, and entrepreneurial support. A survey of returnees found that close to 30 percent of return migrants reported their intentions to look for a job during their stay. Of those who were actually looking for jobs, only two-thirds of them were engaging in paid work. Migrants also tended to be using informal channels rather than formal channels to find jobs, 3 opting for family/friends for job leads rather than government or private agencies. Due to these challenges, an inability to find work was listed as one of the main difficulties confronting former migrants since returning. Many of the migrants who returned to rural areas are unable, or unwilling, to take advantage of work opportunities in urban areas, and in Chisinau in particular, because the remuneration is not sufficiently high to defray the costs of transportation and lodging. Meanwhile, very few migrants intend to invest the funds and knowledge earner abroad in business opportunities in Moldova. This reluctance to invest in a business in Moldova is driven by a number of factors, including corruption, as well as not having anyone to defend their interests in front of the authorities, high bureaucracy / reporting burden, or too many controls. Lack of own financial resources and difficulties in accessing bank loans, as well as high taxes are also common deterrents. While the Government has a number of incentive programs aiming to help integrate the diaspora into the labor market and local community development, return migrants report limited knowledge of their existence. Providing more effective information and assistance to help reintegrate return migrants, and easing qualifying requirements for unemployment benefits could help to offset some of the adverse effects of the COVID-19 crisis on returnee households particularly in the context of a challenging and unfamiliar labor market. Over the short-to-medium term, the majority of return migrants plan to return to their countries of destination, primarily on account of unfavorable assessment of work opportunities and broader socio- economic conditions in Moldova, as well as high reservation wages. Many migrants note improving living conditions, better employment opportunities and a reduction in corruptions among key factors that would influence their decision to return to (or remain in) Moldova permanently. In particular, the median reported monthly wage that would motivate a migrant to return to Moldova to seek employment is, at MDL 15,000, almost double the net monthly wage in Moldova in 2020. Wage requirements for those with higher levels of education, those from the capital city, and those with work experience in EU countries are higher, on average. While return migrants should be supported, the welfare of those who remained abroad should not be overlooked. The resilience of remittances seen during this period are also a testimony to the hard work and sacrifices made by migrants abroad to continue supporting their families back home. Yet, the pandemic has left migrants vulnerable to joblessness, abuse or breach of contract by employers, as well as risks of contagion. Aside from humanitarian considerations, providing migrants access to adequate housing and health care is critical in keeping host communities safe from the pandemic. Some have also been left stranded, unable to leave due to a combination of limited finances, travel restrictions or compromised legal status. Continued support needs to be extended to these migrants by the Moldovan government through its embassies and consulates abroad and where possible, in partnership with the host governments. 4 I. Introduction The COVID-19 pandemic has had a devastating effect on the economy and livelihoods in Moldova. According to data as of the end of November 2020, more than 120,000 people have been infected with the novel coronavirus, and more than 2,500 have succumbed to the virus. As a share of population, Moldova, along with the United States and some other countries in Europe and Latin America, have been among the most severely affected countries worldwide (See Annex Error! Reference source not found.A1). The Moldovan economy is projected to contract by 5.2 percent in 2020, in contrast to a pre- crisis growth projection of 3.6 percent as a result of pandemic containment measures undertaken by both the governments of Moldova and of its trade partners, and the corresponding fall in the aggregate demand. While the economy is envisaged to recover in 2021, (with the latest annual growth forecast of 3.5 percent), significant uncertainty remains which is expected to put a drag on economic activity and keep growth below potential over the medium term (World Bank, 2020b). The pandemic along with the mitigating measures put in place, have also had significant welfare impacts on households in Moldova. According to data from a special COVID-19 module of a nationally representative Household Budget Survey for the second quarter of 2020, 17 percent of sampled households in Moldova reported either reductions or loss of employment income, and 20 percent of households have had to reduce food expenditures to cope with the effects of the pandemic. Beyond monetary losses, 8 percent of households stated difficulties in accessing necessary medical services, while families with children experienced significant difficulties in accessing online education services. Disturbingly, almost 40 percent of the population reported experiencing depression, stress and/or anxiety on account of the pandemic. Migration and remittances could be among the key pathways through which the pandemic impacts households in Moldova. As detailed in the next section, a high share of Moldovans work abroad and a quarter of households in Moldova rely on remittances as a key source of income. The simultaneous disruption of economic activity, both at home and in migrants’ destination countries can limit the capacity of remittances to insure households in Moldova from exogenous shocks. The data from the COVID-19 HBS module revealed that 15 percent of households with migrant members abroad had a returnee on account of COVID-19-related job loss, while 8.3 percent of such households reported the loss of remittances from abroad. The purpose of this paper is to quantify the labor, migration and remittances channels through which COVID-19 pandemic may affect household welfare. Specifically, the paper estimates using the 2018 HBS for Moldova, which is a nationally representative household survey data, the poverty impact of various scenarios linked to job losses, reduction or cessation in remittances from abroad and the return of migrants to Moldova. The methodology accounts for the specific patterns of Moldova migrant and remittance flows from different countries, as well as the expected evolution of key labor market indicators throughout the pandemic. In addition, the paper attempts to estimate the distributional impact of some of the mitigating social protection measures that were deployed by the Government of Moldova (GOM). The rest of the paper is organized as follows. The next section briefly describes the scale of migration and remittances in Moldova, including the reliance on households on remittances with respect to the overall 5 household incomes. Section 3 presents an overview of the main dimensions of the impact of COVID-19 in Moldova, including those related to migration and remittances, in order to frame the simulations that follow. Section 4 describes the data used in the analysis and the main assumption for the simulation models. Section 5 reports the main estimates of poverty impacts of COVID-19, abstracting from policy response measures. Section 6 then accounts for some of the key mitigating measures and estimates the degree to which these attenuate the poverty impact of the pandemic. Section 7 considers some of the longer-term impacts of the pandemic, including the implications of re-migrating of (a share of) returnees back to their pre-pandemic destination countries. Section 8 examines the labor market experience of return migrants. Section 9 provides concluding remarks. II. Migration and remittances as a key livelihood source in Moldova A high emigration rate is one of the defining features of Moldova’s post-Soviet transition. According to official data on border crossings, some 764,000 Moldovan citizens had migrated abroad by the end of 2016 – more than a quarter of the stable population of Moldova according to the Population Census of 2014. 1 Some 300,000 of these migrants were overseas for a period longer than 12 months (Ministerul Afacerilor Interne al Republicii Moldova, 2017). The primary destination countries for Moldovans are the Russian Federation (about 56 percent of the migrant stock), followed by Europe (with 30 percent of stock where Italy has dominated, followed by Poland and Romania and now increasingly Germany). Men typically migrate to Russia for construction-related jobs while a larger share of women have moved to Europe, particularly to Italy for jobs in care-giving or domestic work. A recent International Organization for Migration (IOM) COVID-19 rapid assessment study entitled “Impact of COVID-19 on mobility of Moldovan migrants” (hereafter IOM migrant and remittance survey) capturing migration and remittance dynamics during the pandemic finds that 36.4 percent of migrants reported construction as the sector of activity of their last trip while 26.6 percent reported the household sector/homecare as their sector of work.2 The corollary of high emigration rate is the high inflow of remittances to Moldova. According to World Bank data, Moldova received an estimated US$1.87 billion in 2019, which has increasingly exceeded foreign direct investment, official development aid flows or portfolio inflows (see Figure 2). With remittances being the equivalent of 15.6 percent of GDP in remittances in 2019– the 13th largest value in the world - only Tajikistan and Kyrgyz Republic have seen higher inflows among countries in the Europe and Central Asia region at around 30 percent of GDP in 2019 (See Figure 3). 1 Data from Moldovan diplomatic missions abroad placed the number of Moldovan citizens residing overseas at even higher, at approximately 805,000 in 2015 (Ministerul Afacerilor Interne al Republicii Moldova, 2017). 2 The IOM phone survey which was done in collaboration with the World Bank, sampled both return migrants and migrant households. The household survey consisted of 2004 households that currently have family members who left for work or family members had returned from abroad where they have stayed with the purpose of employment during the last 6 months. The return migrant survey consisted of 1018 individuals who were abroad for work for at least 6 months prior. The data was collected over the period August-September 2020. 6 Figure 2: Remittances, FDI, portfolio and ODA Figure 3: Remittances as a share of GDP in 2019 – inflows to Moldova (2000-2019) ECA region (Percentage of GDP, 2019e) Remittances FDI 29.7 29.6 Portfolio flows ODA ($ million) 2,400 15.6 15.1 2,100 12.3 11.9 11.8 10.5 10.4 9.3 1,800 1,500 1,200 900 600 300 0 -300 Source: World Bank staff estimates using World Bank KNOMAD data and WDI. At the micro level, the 2018 HBS finds that more than 25 percent of households receive international remittances, accounting for approximately 30 percent of their income.3 Remittances are largely a rural phenomenon in Moldova —of the households that receive remittances, 63.4 percent are located in rural areas. More specifically, 28 percent are in the Nord region, 32 percent are in the Centru region, 23 percent are in the Sud region which are all rural and only 17 percent, the lowest proportion reside in the urban Chisinau region. On average, households receive MDL 28,452 or US$3,545 annually in remittances and such receipts are typically pro-poor.4 Close to 70 percent of migrants come from households in rural areas. The majority of migrants are long term migrants with 66 percent staying more than 12 months in their destination country, 16 percent staying between 6-12 months and 19 percent residing less than 6 months. The most common migrant destinations are Russia (37 percent), and Italy (24 percent). Migrants from Russia, Italy and “Other countries” comprise the majority of migrants over all three migration durations— less than 6 months, 6-12 months and over 12 months. The 2016 World Bank Systematic Country Diagnostic (SCD) for Moldova highlighted that remittances are one of the primary drivers of poverty reduction in Moldova in recent years. After pensions, growth in remittances is the second most important source of disposable income growth during the 2011-2016 window (Cojocaru and Matytsin, 2018 and Figure 4). Remittance growth was more prominent at the bottom of the income distribution (particularly for the bottom quintile), consistent with its poverty reducing effect. Remittances during the 2011-2016 period also grew relatively faster in rural areas, both in the Bottom 40 (B40) and the Top 60 (T60) income groups (Figure 5). 3 As a corollary to a quarter of households receiving remittances, 12 percent of households reported having at least one household member working abroad. 4 World Bank, Moldova Poverty Assessment 2016. 7 Figure 4: Shapley decomposition of disposable Figure 5: Shapley decomposition of disposable income growth 2011-2016, by quintile income growth 2011-2016, by urban/rural labor income labor income agricultural income agricultural income pensions pensions social assistance social assistance remittances remittances other income other income Total Total -1.0 0.0 1.0 2.0 3.0 -0.5 0.5 1.5 2.5 3.5 1 2 3 4 5 rural b40 rural t60 urban b40 urban t60 Source: Cojocaru and Matytsin (2018). Notes: Estimates based on HBS data. The high rate of emigration is also a contributing factor to the country’s rapidly aging profile. Among long- term migrants (over 12 months abroad) in 2016, only 6 percent were 55 years of age or older, and the bulk of migrants, seasonal or longer-term, were in the 20-39 age cohort (ILO, 2017). When combined with a low total fertility rate (1.24 in 2014, one of the lowest in ECA) and relatively low life expectancy, Moldova is at the top of the list of ECA countries, ranked by the severity of challenges presented by population aging (Bussolo, Koettl and Sinnott, 2015). High outward migration and a low fertility rate is expected to cause the working-age population to decline by 0.8 percent annually in the next two decades and contribute to a trebling of the old-age dependency ratio to 45.7 percent (World Bank, 2017). A high old- age dependency ratio puts upward fiscal pressure on pension systems as the pool of working-age individuals to generate tax revenue shrinks, though private transfers from abroad may help compensate families directly. Meanwhile, current migration patterns are also resulting in children and elderly being left behind. In 2015, some 41,000 children (6 percent of total in the 0-17 years age group) were left without at least one parent on account of emigration, with grandparents often assuming the role of caretakers (IOM, 2016). III. Key channels of welfare impacts of COVID-19 pandemic The COVID-19 pandemic can potentially affect household welfare through a variety of channels, summarized in Figure 6. Direct shocks to labor income, either through reduced or lost employment and earnings are a key channel of impact, as labor earnings commonly accounts for a large share of overall disposable income for many households. However, household welfare can also be impacted through non- labor income channels, especially for households relying heavily on remittances or safety nets. In addition to these channels affecting labor and non-labor income, the pandemic can result in price changes for essential goods and services, including health-care expenditures, which may force credit-constrained households to restrict their consumption of essential goods and services. Finally, broader, non-monetary, 8 dimensions of welfare, can also be affected by the disruption of basic services such as health and education. This paper focuses on the first two channels, specifically those operating through the labor market channel and via the remittance dimension of non-labor income, given its importance in Moldova. Social protection measures deployed by the GOM to mitigate the impacts of the COVID-19 pandemic will be explored through the safety-nets component of the non-labor income pathway. The analysis is partial as it abstracts from the price effects and broader non-monetary dimensions of welfare. It does, however, consider the main pathways that would be relevant for assessing the impacts of the pandemic on monetary poverty, which is the goal of this paper. Figure 6: Conceptual framework for identifying COVID-19 impacts on household welfare Direct: Lost earnings due to illness Labor Income Indirect: earnings /employment shocks Remittances and private transfers Non-Labor income Public transfers Welfare (monetary & non-monetary) Price changes Direct effect on consumption Out of pocket costs of health care Saturation of health system (NCD, etc.) Service disruptions Schools: nutrition, learning, dropouts Source: World Bank (2020) The labor market channel In response to the coronavirus crisis, the GOM instituted a state of emergency on March 17th and implemented containment measures aimed at reducing the spread of the virus. Among the measures introduced were the closure of all educational institutions and several public venues, including gyms, museums, and theaters, bars and restaurants. Strict transportation restrictions were instituted, including the suspension of air and rail traffic, as well as the closure of 70 out of Moldova’s 81 land border crossings with Romania and Ukraine. Additional quarantine measures included restricted opening and working hours for establishments, prohibition of in-person meetings, public events and other mass events; 9 requiring that schools and universities shift to online and distance-learning methods; and the temporary suspension of courts processing criminal and administrative cases. To a certain extent, the employment structure of Moldova insulates the country from some of the adverse labor income effects of the COVID-19 crisis (See Annex Figure A2). Employment in public administration, education, health and social work accounts for the greatest proportion of employment at 22.6 percent followed by 21.0 percent in agriculture, forestry and fishery. Based on a high probability that jobs in the public administration, education, health and social work sector would be deemed essential and that the majority of work in the agricultural sector is rural self-employment, these sectors are expected to be largely shielded from the impacts of the crisis.5 However, Moldova has a high proportion of employment in service-oriented sectors such as trade, hotels and restaurants which are particularly vulnerable to lockdowns. The containment measures introduced had adverse impacts on the labor market as the economy grounded to a halt. According to the country’s Labor Force Survey data, at the height of the crisis in Q2, the number of employed people fell by 79,600 or 8.8 percent compared to the same quarter in 2019. While the unemployment rate was not affected, broader labor market measure of underemployment increased significantly in Q2. For instance, the LU2 indicator for Q2, which combines the unemployment rate and time-related underemployment for individuals who are working but would like and are available to work more hours, increased by 3.5 percentage points year-on-year. According to data from the LFS survey, 200,000 individuals, or a quarter of those employed in Q2 of 2020 reported that their jobs were affected by COVID-19. The labor market disruption associated with the lockdown were not uniform across sectors. Preliminary estimates from the National Bureau of Statistics of Moldova show that in Q2 2020 which is at the height of the crisis, the greatest decline in employment relative to the same quarter in 2019 was in the service sector (at 10.6 percent), followed by construction (9.8 percent), industry (7.8 percent) and with agriculture falling the least (at 4.7 percent).6 The impact of the pandemic on the services sector is corroborated by financial reports from the State Tax Service for May 2020 when compared to February 2020: some 32 thousand employees in the real sector of the economy lost their jobs, with travel agencies (a fall of 89 percent), and the hotel, restaurants and catering (HoReCa) sector (down by 44.6 percent) being by far the most affected. These two sectors experienced a sharp decline in sales in March-May 2020 compared to the same period in 2019.7 By comparison, in the food industry and in the construction sectors, the number of employees fell by about 4 percent over the same period (Government of Moldova, 2020). The volume of production also increased in the food industry and in the construction sector, in part on account of the returning migrants engaging in renovations or repairs activities (Government of Moldova, 2020). 5 According the Food and Agricultural Organization, 25% of those employed in agriculture work as employees in agri-enterprises and 75% of those employed are classes as self-employed. http://www.fao.org/family- farming/countries/mda/en/#:~:text=The%20agriculture%20sector%20employs%20more,are%20classified%20as%20self%2Demployed. 6 https://statistica.gov.md/newsview.php?l=en&idc=168&id=6681 7 Sales fell by 88.6 percent for travel agencies while it was down by 67 percent for the HoReCa sector. 10 Fig 7: Employment and Unemployment by Fig 8: Change in employment rate and quarter, Q1 2019-Q3 2020 unemployment rate by quarter 920 80 14.0 900 70 12.0 Number of Unemployed Number of Employed 880 60 10.0 860 50 840 8.0 40 6.0 820 800 30 4.0 780 20 2.0 760 10 0.0 740 0 I II III IV I II III Q1 Q2 Q3 Q4 Q1 Q2 Q3 2019 2019 2019 2019 2020 2020 2020 2019 2020 Employed Unemployed LU1 LU2 Note: LU1 – unemployment rate; LU2 – combined rate of time-related underemployment and unemployment Source: Official estimates based on LFS data. With the easing of containment measures and the resumption of economic activity, key labor market indicators improved in Q3 of 2020 with the employment rate lower by 1.7 percentage points than in the corresponding quarter of 2019. The measure of labor market underemployment also fell in Q3 of 2020 to levels below that of Q3 2019. The numbers of those reporting in Q3 that their jobs were affected by the COVID-19 pandemic came down to 21,000, or 2.5 percent of the total occupied population, of which more than two thirds of these were from urban areas. Among those affected by the pandemic, a third reported not working at all, while almost half reported working fewer hours. The migration / remittances channel The heavy reliance of Moldovan households on remittances as a key source of livelihood is very important, as detailed in the previous chapter. Earlier studies have highlighted the counter-cyclical nature of remittances, which can serve as an insurance for recipient households against adverse economic shocks (Frankel, 2009). However, Moldova’s circumstances may indicate otherwise. The landlocked country is a small open economy and recessions in key trading partners such as the Russian Federation and the European Union, which are also the primary hosts of Moldovan migrants, tend to move in tandem with deteriorating economic performance in Moldova (e.g. recessions of 2009 and 2015). As such, during a region-wide crisis, Moldovan migrants residing abroad have a limited capacity to send funds home precisely at the time when their households back home and the general economy, would benefit from remittances the most. Figure 9 shows that remittances to Moldova declined during both recent recession episodes in Russia and Moldova. 11 Figure 9: Growth rate of remittances to Moldova vis-à-vis real GDP growth in Moldova and Russia (Growth rate, %) 60 Remittances to Moldova 15 50 Russia GDP 40 Moldova GDP 10 30 20 5 10 0 -10 0 -20 -30 -5 -40 -50 -10 2007 2000 2001 2002 2003 2004 2005 2006 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Source: World Bank staff estimates using WDI and WB Commodity Price data. According to the October 2020 World Economic Outlook estimates, the Italian economy is expected to contract at the rate of 10.6 percent in 2020, with Russian GDP growth projected at negative 4.1 percent. Given that Russia and Italy together account for approximately 80 percent of Moldova’s migrant stock8, these dynamics affect both return migration and remittances. According to findings from a previous IOM diaspora COVID-19 rapid field assessment report conducted in March, 41 percent of migrants indicated that they had stopped sending remittances all-together, 19 percent of migrants reported that the amount of remittances sent had decreased by 71-90 percent and a further 18 percent of all migrants reported that they would cease sending remittances in the near future due to the impact of COVID-19.9 At the national level, the volume of remittances in Moldova fell by some US52.4 million or 11.8 percent in Q2 2020 over Q1 as the crisis reached its peak but is projected to show some recovery in Q3 (Figure 10). Compared with the corresponding quarter of 2019, remittance inflows fell by 15.2 percent or 1.1 percent of GDP. This decline is likely to have had detrimental impacts on households’ purchasing power given the high share of households that are dependent on remittances. Fig 10: Evolution of remittances flows Fig 11: Quarterly change in remittances 600 19 10.0 6.7 18 5.0 550 0.7 1.9 17 0.0 500 16 0.0 15 -5.0 -1.1 450 14 -10.0 400 13 -15.0 350 12 -15.3 Q1 Q2 Q3 Q4 Q1 Q2 Q3 -20.0 2019 2019 2019 2019 2020 2020 2020* ΔQ1 ΔQ2 ΔQ3* Remittance Inflows Remittances as a % of GDP Remittance Inflows Remittances as a % of GDP Source: National Bank of Moldova. *denotes estimated numbers. 8https://rovienna.iom.int/news/iom-joins-call-better-access-remittances-moldova 9IOM rapid field assessment of the impact of COVID-19 on the wellbeing of the Moldovan diaspora: an evidence base regarding migrants’ coping strategies and contributions, IOM(2020). The interviews were conducted by phone. 12 Reduced economic activity in destination countries, either due to the direct effects of mitigation measures or on account of reduced aggregate demand, can not only lead to a reduction in the flow of remittances, but also to a larger flow of return migrants. A Rapid Diaspora Survey from IOM conducted in April-May 2020 across 10 destination countries revealed some 30 percent of those surveyed having plans to return to Moldova, an equivalent of 255,000 returnees (IOM, 2020), raising concerns that unemployment could rise sharply and safety nets could become strained.10 With the second wave of the crisis in Europe in the Fall of 2020, this trend may continue into 2021 as the economic conditions of destination countries continue to worsen or recovery stalls. This analysis below aims to shed some light on the degree to which these concerns have come to fruiting throughout 2020 and provide estimates of the magnitudes of the welfare implications of changes in remittances flows and return migration. IV. Data description and empirical strategy To quantify some of the welfare effects of COVID-19, this paper presents a set of simulations drawing on a nationally-representative household survey data and taking into account the dynamics of key economic, labor market and migration indicators through 2020, in addition to the corresponding government responses over a 12-month period. The simulations are described more fully below, but as noted above, they capture only a subset of the potential effects of COVID-19; notable exclusions are price effects, health effects, and second-order demand effects due to losses in employment and income. Furthermore, the effects evaluated in this paper do not account for potential behavioral responses to economic shocks. Lastly, the simulations are reflective of the data available at the time and its limitation and should be interpreted in that context. Data The main source of data for this analysis is Moldova’s 2018 Household Budget Survey (HBS) which is a nationally representative household survey conducted annually that captures information on income, expenses, and consumption in addition to demographic and social indicators. The 2018 HBS covered 5,142 households containing 11,355 individuals. The measure of poverty used is expenditure-based and is defined as the proportion of individuals who have consumption expenditure less than the upper middle- income line of $US5.50 2011 PPP per day or MDL 43.80. The parameters for the simulations are derived from quarterly labor force indicators produced by The National Bureau of Statistics of Moldova (NBS) and select indicators from the IOM survey of households with migrants and returned migrants (see below). The analysis also relies on the survey of household with migrants and of return migrants, undertaken by CBS Research in collaboration with IOM and the World Bank. The data from this survey is used both to inform the input parameters of the simulations, as well as to gain additional insights into the experience and expectations of migrants throughout the pandemic. The survey has 2 components: (i) a stratified, nationally representative sample of 2,004 households that currently have family members abroad, or family members who have returned from abroad where they have been for work during the last 6 months; and (ii) a quota sample of 1,018 migrants, who have returned from work abroad during the past 6 months, 10 For more details: https://www.iom.int/news/iom-un-migration-agency-mission-moldova-issues-first-rapid- diaspora-survey-report-looking 13 with the quotas based on administrative data from the Border Police of the Republic of Moldova on the stock of entries and exits in/from Moldova by the citizens of Moldova. The survey data was collected during July-September 2020. In addition, the study also draws on qualitative information from focus groups as well as in-depth interviews with return migrants and family members with current migrants undertaken during September-October 2020, with a total of 39 family members, 20 returned migrants and 6 migrants who left during the COVID-19 pandemic having been interviewed (Cantarji et al., 2021). Simulations setup Baseline Scenario The baseline case corresponds to the poverty rate and poverty gap ratio based on a monthly consumption per capita in Moldova in 2018 and corresponding poverty line of MDL 1332.34, the lei equivalent of the US$5.50 PPP per day or MDL 43.80. At baseline, the headcount poverty rate is 12.9 percent or approximately 429,000 people while the poverty gap is 2.3 percent. In departing from the baseline, we account for three separate channels through which households are impacted by the pandemic. A. Labor Channel In the labor simulation we consider employment losses as a result of the COVID-19 pandemic. Absent data on the proportion of households that experienced income declines and its magnitude, we are unable to simulate income declines. Given this limitation, the employment impacts are likely to be understated and these impacts should be interpreted as a lower bound for impacts in the labor market. We aggregate sectors of employment into four categories—agriculture, industry, construction and services—and we use the employment losses as reported by the National Bureau of Statistics for 2020 as compared with 2019 to parameterize the model.11 We randomly assign individuals out of employment based on their sector of employment and the corresponding sectoral declines in employment based on the labor force survey. The declines in employment are as follows: Agriculture, 3.8 percent; Industry, 5.2 percent; Construction, 1.6 percent; and Services, 4.7 percent. Based on these proportions, we estimate the number of employed to decline by 45,050, a figure slightly higher than the 38,200 annual decline in employment registered for 2020. Nevertheless, caution should be taken in interpreting the simulation results as the HBS is not designed to be representative of the Moldovan labor force and a substantial proportion of individuals report employment but zero wages.12 B. Remittances Channel In the remittance simulation, we consider the impact of changes in remittances on household welfare. According to data from the IOM survey of households with migrants, some 25 percent of households reported the volume of remittances to have declined during the pandemic, either significantly or 11Aggregation was done in keeping with the International Standard Industrial Classification of All Economic Activities (ISIC) by the International Labor Organization. The classifications can be found at https://ilostat.ilo.org/resources/methods/classification-economic-activities/ 12In the HBS, the proportion of employment is 53% in Agriculture, 10% in Industry, 4% in Construction and 33% in Services. In comparison, the official labor force statistics show 19% employment in Agriculture, 15% employment in Industry, 5.3% employment in Construction and 60.7% employment in services. For individuals who report non-zero wages, the proportions are 14% in Agriculture, 20% in Industry, 3% in Construction and 63% in services. 14 insignificantly, while about 20 percent of households reported an increase in the volume of remittances and 38 percent reported the volume of remittances to have remained unchanged (Cantarji et al., 2020). The evolution of remittances during the pandemic is important to household welfare. Overall, among households with migrants about 15 percent report spending less on food, and 12 percent report spending less on health compared to the same period during the prior year before the pandemic; about 20 percent report having either sufficient or insignificant savings. However, the incidence of coping mechanisms such as reducing consumption of necessities such as food and health expenditures, as well as availability of savings, is notably lower among households reporting that remittances decreased during the pandemic. Figure 12: Share of households that reduced food Figure 13: Share of households reporting and health expenditures, by remittances availability of savings, by remittances dynamics dynamics category category 25.0% 40.0% 35.0% 20.0% 30.0% 15.0% 25.0% 20.0% 10.0% 15.0% 10.0% 5.0% 5.0% 0.0% 0.0% Increase in the Without Decrease in the Increase in the Without Decrease in the volume of changes volume of volume of changes volume of remittances remittances remittances remittances Source: Cantarji et al. (2020) based on data from the IOM migrant and remittance survey. However, in order to simulate the impact of changes in remittances on monetary poverty, we need to simulate the pandemic-related remittances dynamics in the HBS survey data. Guided by the IOM diaspora survey13, we disaggregate the proportion of households reporting declines and increases in remittances based on the country where the migrant is residing.14 Respondents were asked to estimate how the amount of remittances has changed given the following responses: “has increased significantly”, “has increased but insignificantly”, “hasn’t changed”, “has dropped but insignificantly”, “has dropped significantly”, and “I cannot answer”. Absent numerical data on the magnitude of income declines, we interpret “significant” to mean 20 percent and “insignificant” to mean 5 percent in the case of both reports of increases and declines in remittances. In total, 6.6 percent of households reported reduced remittances which is slightly lower than the 8.3 percent of households that reported reduced remittances in the HBS COVID-19 module. These assumptions correspond to a disaggregation of changes in remittances by 13 The Impact of COVID-19 on mobility of Moldovan migrants, IOM(2020). The survey covered 1,186 Moldovan migrants. The proportions surveyed were as follows: Russia (16%), Italy (21%), Israel (3%), Portugal (2%), Ukraine (1%), Turkey (1%), Greece (1%), Other (55%). By contrast, the proportions in the household survey used in this simulation are: Russia (37%), Italy (24%), Israel (5%), Portugal (2%), Ukraine (1%), Turkey (3%), Greece (0%), Other (29%).The IOM survey results should be interpreted in the context of the well-known limitations of online surveys, specifically, the data may exclude the most vulnerable migrants, who may not have access to equipment required to participate in the survey, or migrants with irregular status who in addition may refuse to answer out of fear. Additionally, there may be less familiarity with digital surveys among certain groups of migrants that could affect the quality of the data. 14 For those households with multiple migrants from multiple countries, we use the mode and the minmode. 15 country shown in Table 1.15 In the appendix, we further consider an alternate scenario where we interpret “significant” to mean 33 1/3 percent and “insignificant” to mean 10 percent. We find using this specification the increase over baseline poverty goes from 0.2 to 0.4 in the remittances scenario and 2.7 to 2.9 in the “All” scenario. It is important to note that the remittances channel captures the effect of migrants who have remained abroad but who have altered their remittance sending behaviour. Table 1: Change in remittances by category by country Country Significant Insignificant Significant Insignificant Hasn’t I can’t Decrease Decrease Increase Increase Changed answer (-20%) (-5%) (+20%) (+5%) Russia 16 8 12 12 32 20 Italy 15 13 6 11 41 14 Portugal 7 14 7 0 56 16 Turkey 0 0 0 0 100 0 Ukraine 26 22 23 0 0 29 Israel 13 5 5 11 50 16 Greece 22 0 0 0 22 56 Other 11 15 14 8 39 16 Source: Staff calculations using data from the IOM migrant and remittance survey (2020). C. Return Migrant Channel Via the return migrant channel, we consider the impact of migrants returning home from their destination countries due to the fallout from COVID-19. According to Border Police data, just under 316,000 Moldovan migrants returned over the period January-September 2020 with over three-quarters of them returning over the period March-September.16 Notably, the number of returning migrants throughout the period March - September 2020 is considerably lower than the corresponding number of returnees during the same months of 2018 and 2019, consistent with restrictions in mobility imposed across countries due to the pandemic. 15 We average the remittance changes of the following countries to get the return rate for other countries: Spain, France, Germany, the United Kingdom, Belgium, Romania, Czech Republic, Ireland, Netherlands and Poland. 16 Number of State border crossings by nationals of the Republic of Moldova who entered the country after an absence more than 90 consecutive days (Border Police). 16 Figure 14: Number of returned migrants, monthly 140.0 120.0 100.0 80.0 58.3 60.0 53.9 38.8 42.2 36.2 34.1 40.0 20.2 21.8 20.0 10.1 0.0 JANUARY FEBRUARY MARCH APRILIE MAY JUNE JULY AUGUST SEPTEMBER 2018 2019 2020 Note: Number of crossings of the State Border by citizens of the Republic of Moldova who entered the country after an absence of more than 90 consecutive days. Source: Cantarji et al. (2020) based on administrative data from the Border Police. Furthermore, it should be noted that the number of entries into Moldova during this period is only somewhat larger, and follows the same dynamics over time, as the number of exits from Moldova. Overall, between January 18 and September 27, 2020, there were 2.35 million crossings into the Republic of Moldova, and 2.29 million crossings out of Moldova, with roughly 60,000 net “return migrants”. Figure 15: Flow of entries and exits in/from Moldova during March-September 2020 200000 180000 160000 140000 120000 100000 80000 60000 40000 20000 0 Entries Exits Source: Cantarji et al. (2020) based on administrative data from the Border Police. 17 Results from IOM migrant and remittance survey estimates that 23 percent of households with migrant members have had a migrant who returned home from abroad in 2020. Return migrants tend to be disproportionately male, are on average 30-39 years and have vocational education and are largely in the rural and Centre regions. The returnees were primarily from Italy (18 percent), Russia (15 percent), Germany (13 percent), France (11 percent) and the United Kingdom (11 percent). The report confirms that individuals with limited migration spells abroad were most likely to return with 85 percent of those who returned being abroad for less than a year. These migrants tend to be disproportionately employed in the construction and domestic care sectors. Interestingly, only 22 percent of migrants reported returning for reasons specifically ties to COVID-19 and only 10 percent for reasons potentially related to COVID-19, though it is likely that this figure is understated. Not all returns need be attributable to the pandemic; some migrants may be returning as previously planned, or for other family reasons unrelated to the COVID-19 pandemic. We rely on the data from the IOM survey to isolate the returns that are COVID-19 related. In particular, we assume the following reasons for return are attributed to COVID-19 related impact based on the migrants’ specified reasons for return: “The number of persons infected by COVID19 was increasing in the country where he/she last stayed ”, “Lost the job because of the pandemic, “Is on temporary leave from job abroad due to COVID19 pandemic”, “The wage abroad has been reduced”, “His/her working hours have been reduced”, “The attitude of the residents of the country where he/she was staying towards migrants has worsened because of the pandemic“, and “Forced to leave because of the pandemic by the authorities of the country where he/she was staying”. According to a survey of return migrants, roughly 30 percent of return migrants came back for reasons either directly or potentially attributable to COVID-19, which would amount to roughly 18,000 return migrants out of the total 60,000 net return migrants based on Border Police data. Since the IOM migrant and remittance survey was fielded in July through to August, we further limit the data to individuals who have returned in the past 2-3 months to coincide with onset of the pandemic. After these adjustments, the return rate by country is as follows: Russia (6 percent), Italy (7 percent), Portugal (6 percent), Turkey (9 percent), Ukraine (4 percent), Israel (1 percent), Greece (10 percent), and Other Countries (12 percent).17 Combined we have 15,650 migrants returning on account of COVID-19, out of a potential 211,000 migrants returning, which is close to the aforementioned estimates based on Border Police and IOM survey data. We further assume that 70 percent of returned migrants return with savings equivalent to their annual remittances based on the proportion of returned migrants that report “savings/money raised from work carried out abroad” as one of their sources of income.18 The savings amount to MDL 357,000 annually or just under 1 percent of the total amount of remittances in 2018. In contrast to the remittance channel, the return migrant channel captures the impact of migrants who return from abroad under certain assumption on the amounts of money they return with. All Scenario: 17 We average the return rate of the following countries to get the return rate for other countries: Spain, France, Germany, the United Kingdom, Belgium, Romania, Czech Republic, Ireland, Netherlands and Poland. 18 We consider the return migrant scenario as distinct from the remittances scenario as the remittance question was fielded to households who currently had a migrant abroad. 18 In the “All scenario”, we combine the three previously mentioned channels, allowing them to simultaneously affect household income and consumption.19 V. Main results: unmitigated effects of the pandemic on poverty and the profile of the new poor The COVID-19 pandemic has affected the livelihoods of many households in Moldova. According to a special COVID-19 module of the 2020 Household Budget Survey, in the second quarter of 2020 17 percent of households reported a reduction or loss of incomes from employment, 8.3 percent of households reported a reduction or loss of remittances from abroad (during the 3rd quarter of 2020 the same was reported by 12.1 percent and 4.7 percent of households respectively). Moreover, 1 in 5 households had to reduce expenditures on food during the 2nd quarter and 4.6 percent of households had to resort to partial payment or non-payment of utility bills (12 percent and 3 percent respectively in the 3rd quarter). Similarly, 1 in 5 households had to rely on savings to cover household expenditures during the 2nd quarter of 2020.20 In the survey of households with migrants, 26 percent reported that the COVID-19 pandemic has affected the well-being of their households very much. Consistent with estimates from the HBS COVID-19 survey, 15 percent of households in the survey of households with migrants reported having to reduce expenditures on food in comparison to the pre-pandemic spending levels, and 5 percent of households had to reduce spending on utilities (Cantarji et al., 2020). This section supplements the evidence on broader welfare impacts of the pandemic on income and expenditure patters, as well as subjective measures of well-being during the pandemic, with simulations that aim to estimate the impact of the COVID-19 pandemic on measures of monetary poverty in Moldova, and on the profile of the population below the poverty line. Estimated impacts of the pandemic on the incidence and depth of poverty The simulations suggest that in the absence of direct COVID-19 related social interventions to workers, return migrants and households, the poverty rate is likely to rise. If the remittances, labor and migration channels are taken into account together, poverty would rise by 2.7 percentage points from 12.9 percent to 15.5 percent. The labor channel has the greatest effect on the poverty rate, resulting in a 1.6 percentage points increase from the baseline, however, this effect is likely to be understated as it only accounts for employment losses and does not capture income declines By contrast, the remittance and return migrant simulations , lead to increases of only 0.2, and 0.7 percentage points, respectively. The relatively muted impact of these scenarios is attributable to a number of factors. First, a relatively low proportion of households reported declines in remittances and some even reported increases in remittances. Second, remittances tend go to relatively well-off households which may have additional buffers to prevent them from falling into poverty. Last, a relatively small proportion of migrants returned, the majority of which return with savings, limiting the impact of return migration to those returning without savings or those returning with inadequate savings becoming net dependents in the household. 19Note that the “All simulation” will not necessary be equivalent to the sum of the individual simulations due to potential overlaps in the number of individuals classed as poor across simulations. 20 For details, see: https://statistica.gov.md/newsview.php?l=ro&idc=168&id=6773 19 Analyzing different subcategories of households reveals stronger impacts. Among remittance receiving households, the poverty rate rises 0.8 percentage points from 11.2 percent in baseline to 12.0 under the remittances scenario. Among female-headed households, under the “All” scenario baseline poverty increases from 11.0 percent to 13.6 percent. For male-headed households, the commensurate increase is from 13.9 percent to 16.6 percent. Much of the increase in baseline poverty among female-headed households is driven by the employment channel which contributed to 1.4 out of the 2.6 percentage point increase, and is likely due to the concentration of women in the services sector which suffered the greatest proportionate decline in employment. Changes in the national poverty gap largely mirror changes in the national headcount poverty rate—the labor channel is responsible for the greatest absolute increase in the poverty gap while the remittance and return migrant channels have relatively muted effects (Figure 17). Figure 16: Simulated effects on the national Figure 17: Simulated effects on the national poverty headcount across various COVID-19 poverty gap across various COVID-19 channels channels relative to the baseline scenario relative to the baseline scenario 1 3.0 Percentage point change over baseline scenario 2.7 1.6 2.5 Change over baseline poverty gap 1.4 1.2 1.2 2.0 1.0 1.6 1.0 1.5 0.8 1.0 0.6 0.7 0.4 0.5 0.2 0.2 0.2 0.0 0.0 0.0 Remit Labor Migrant All Remit Labor Migrant All Source: Author’s estimates based on HBS 2018 data. Results disaggregated at the urban/rural level show that under the “All” scenario, urban poverty would increase by 2.2 percentage points up from 6.1 percent to 8.3 percent whereas rural poverty would increase by 3.0 percentage points up from 18.1 percent to 21.1 percent. The results at the urban/rural level suggest that of the individual scenarios, the labor channel has the greatest effect on the increase in the headcount poverty rate for both urban and rural areas accounting for increases in baseline poverty of 1.8 and 1.5 percentage points, respectively. It is important to note that the effect of the labor channel in rural areas is partially attributable to the effect of the agricultural drought experienced in 2020. The larger effect of labor on baseline poverty in urban areas is likely due to the higher concentration of employment in the services and industry sectors in urban areas. In contrast to urban areas where their impact is absent, falling remittances has an effect in rural areas, accounting for a 0.4 percentage point increase in rural poverty, consistent with remittances being largely a rural phenomenon. Similar conclusions can be drawn to explain the stronger effects of return migration in rural areas—of migrants that were simulated to return, close to 90 percent were associated with households in rural areas consistent with migration being a largely rural phenomenon. 20 Figure 18: Simulated effects on urban/rural poverty headcount across various COVID-19 channels relative to the baseline scenario 5.0 Percentage point change over Urban Rural 4.0 3.0 baseline scenario 3.0 2.2 1.8 2.0 1.5 1.1 1.0 0.4 0.2 0.0 0.0 -1.0 Remit Labor Migrant All Source: Authors’ estimates based on HBS 2018 data. Changes in the urban/rural poverty gap also largely mirror changes in the urban/rural headcount poverty rate (Figure 19). The labor channel is responsible for the greatest absolute increase in both the urban and rural poverty gap and the effect of the crisis on the poverty gap is relative stronger in rural areas. Figure 19: Simulated effects on urban/rural poverty gap across various COVID-19 scenarios relative to the baseline scenario 1.4 Urban Rural 1.3 1.2 Percentage point change over 1.2 1.1 1.0 0.9 baseline scenario 0.8 0.6 0.4 0.3 0.2 0.1 0.0 0.0 0.0 Remit Labor Migrant All Source: Authors’ estimates based on HBS 2018 data. Profile of the new vs existing poor In order to inform adequate social protection responses which is discussed in more detail in the next section, it is useful to examine the characteristics of the “new poor”, that is, those who were not previously poor under the pre-COVID baseline scenario but who turned poor under the “All” scenario. Using this definition of the new poor, the new poor is estimated at around 92,550 individuals from 29,900 households. This stands in comparison to the existing poor that is estimated to be around 429,250 individuals from 139,550 households. Analyzing the simulations that lead to the emergence of the new poor show that 62 percent of the new poor are in households where at least one family member has lost 21 a job, 11 percent are in households that reported a decline in remittances and 30 percent are in households where a migrant has returned.21 It is noteworthy that close to half of the new poor had baseline consumption levels within 30 per cent of the poverty line and could be classified as vulnerable prior to the crisis. Furthermore, over a third had consumption between 40 and 200 per cent of the poverty line at baseline (See Appendix Figure A4). Given the relatively larger magnitude of the labor market impact on household welfare, it is useful to discuss the labor market dynamics of the new poor and compare them with the existing poor. While the new poor has a higher proportion of employed individuals compared with the existing poor (82 percent vs 75 percent) (See Appendix Figure A5), they tended to be concentrated in the services sector (47 percent), agriculture sector (34 percent) and the industry sector (18 percent). By contrast, close to three- quarters of the existing poor are concentrated in agriculture (See Appendix Figure A6). The concentration of the new poor in the industry and services sector is largely consistent with these sectors experiencing the greater declines in employment (5.2 percent and 4.7 percent, respectively). Close to a third of the new poor are concentrated in the agriculture sector. However, this concentration may also be due to the confounding effects of the drought that coincided with the harvesting season in Q3 of 2020. Following on from the disproportionate impact of the crisis on the labor market, the working age population (15-64) comprises just over 70 percent of the new poor whereas the child population (0-14) consists just over a fifth and the elderly population close to 5 percent. The proportion of children and the elderly among the existing poor is higher, accounting for close to 30 percent and 16 percent, respectively (See Appendix Figure A7). The new poor and the existing poor have the same household size, averaging 3 members, however, the new poor tend to have fewer children (0-14 y.o) averaging 0.7 children per household compared with 0.8 children for the existing poor. The new poor have a slightly higher proportion of female-headed households at 37 percent compared with 35 percent for the existing poor. The new poor tends to have a higher share of higher educated individuals with close to 45 percent of the new poor being tertiary educated compared with a quarter of the existing poor (See Appendix Figure A8). This difference could be due to the higher share of more educated individuals in sectors that were hit harder by the pandemic, particularly the services sector (Appendix Figure A9). The new poor are also more likely to be employees (69 percent) and less likely to be self-employed (30 percent) as compared with the existing poor who are more likely to be self-employed (62 percent) and less likely to be employees (35 percent) (Appendix Figure A10). The difference in employment type between the new- and existing poor largely stems from the tendency for self-employment in agriculture while employment in industries and services is mostly employee-based (See Appendix Figure A11). Close to 20 percent of the new poor were concentrated in the Chisinau region compared with less than 3 percent of the existing poor. The difference in the geographic concentration of the new and existing poor in the capital region stems from geographic segregation of industries whereby 35 percent of industrial jobs and 43 percent of services jobs are concentrated in the Chisinau region (Appendix Figure A12). By contrast, the Nord, Centru and Sud regions account for 35, 36 and 27 percent of employment in agriculture, respectively (Appendix Figure A12). 21 The figures do not sum to 100 due to overlapping categories. 22 VI. The welfare effects of mitigating policy measures As part of the emergency response to the COVID-19 pandemic, the Moldovan Government implemented a social assistance package to protect the poor and vulnerable for a period of two months, beginning on March 17th and expiring at the end of the state of emergency on May 16th. As part of this package, the unemployment benefit was increased to MDL 2,775 per month for workers who lose their job and potentially extended coverage to the informally employed, the self-employed, and returning migrants through waiving eligibility criteria and modifying existing legislation.22 The Government also increased the Guaranteed Minimum Income (GMI) from MDL 1,100 to MDL 1,300 per month and raised the child benefit adult equivalency coefficient from 50 to 75 percent (from MDL 553i to MDL 975).23 It should be noted that the policy measures simulated here are only a subset of the measures implemented by the GOM in response to the COVID-19 pandemic. Due to data limitations we limit our simulations to the measures the GOM directly targeted to workers, return migrants, and households in response to the impact of the COVID-19 pandemic. Other implemented measures such as tax breaks and tax deferrals targeted to firms could have spillover welfare impacts on households, however, the simulation of these effects is beyond the scope of this paper. To the extent that these measures led to employment retention, they would partially be captured in lower employment changes as captured by the labor scenario, though not explicitly modeled. With these caveats in mind, the results should be interpreted as approximating a partial equilibrium simulation that assesses the direct welfare impact of selective transfers to the aforementioned segments of the population. In order to simulate the government’s response, we assign the 4,050 individuals who are currently receiving the unemployment benefit the difference between the new unemployment benefit of MDL 2,775 /month and the current unemployment benefit and extend unemployment benefits of MDL 2,775 /month to the 45,050 individuals predicted to lose their jobs in the labor simulation as well as the 700 migrants or 4.3% of the 15,800 simulated return migrants in the migrant simulation based on the reports on the proportion of return migrants who have received the unemployment benefit.24 We modify the child coefficient in the GMI benefit formula from 0.5 to 0.75 and compute the effect of the Guaranteed Minimum Income (GMI) on poverty by assigning individuals the difference between their income and the total GMI adult equivalent of MDL 1,300 if their simulated income under the combination of the remittances, migrant, social assistance (including the new unemployment benefit and the existing Ajutor Social benefit) and labor simulations is less than the GMI.25 Based on a prior complementary analysis on the coverage of the GMI, we further assume that a quarter of households who qualify for the GMI benefit receive it. We abstract away from issues with targeting and leakage of the GMI benefit in our analysis. 22 This was done through waiving the mandatory health insurance requirement to access unemployment benefits. This is however valid for individuals who can prove that they have made the contributions to the health social insurance for 12 months in the past 24 months. All the others, including returning migrants, must pay their contributions first before applying for unemployment assistance. The legislation was also amended to allow individuals without prior work record to be included in the allowance category in addition to the self-employed. (UNDP- Social and Economic Impact Assessment of COVID-19 in Republic of Moldova, 2020) 23 Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures “Living paper” version 11 (June 12, 2020), the World Bank 24 According to the IOM migrant and remittance survey, benefit take-up is low among return migrants, with only 6% of return migrants applied for unemployment benefits (and of these 44% received the benefit). Administrative data from ANOFM shows 777 return migrants having been registered as unemployed during Jan-June 2020. 25 We add employment and child benefits as these benefits are counted as income when the social assistance benefit is determined. Additionally, the new GMI threshold can be interpreted as a “top-up” for individuals already receiving the Ajutor Social benefits. 23 We consider three sub-scenarios: the government’s actual social assistance package as was implemented under the initial 2 month emergency period as , a hypothetical scenario in which the government’s initial social assistance package is extended to a longer 4 month emergency period when the majority of restrictions are expected to be eased, and a hypothetical scenario in which the government’s initial social assistance package is extended to a 6 month period to allow for a further two-month buffering period. It should be noted that these are only approximations to the government’s proposed policy intervention and all results should be interpreted in this context. Furthermore, due to data limitations, we do not directly simulate an increase in the coverage of the unemployment benefit to the informally employed. The costs of the various scenarios are shown in the table below. As a reference point, the GOM’s Ajutor Social budget allocation was 202 million MDL and the budget allocation for the amendment to unemployment benefits was 410 million MDL. Table 2: Estimated annual cost of the 2, 4 and 6-month social assistance scenarios 2-month benefits 4-month benefits 6-month benefits Component MDL Percent MDL Annual Percent MDL Annual Percent Annual Cost of total Cost of total Cost of total cost cost cost Job loss 249,900,000 48 499,800,000 50 749,700,000 51 unemployment benefits Increase in current 9,693,000 2 19,387,000 2 29,080,000 2 unemployment benefits Return migrant 7,190,000 1 14,379,000 1 21,569,000 3 unemployment benefits Total unemployment 264,100,000 50 528,200,000 53 792,300,000 54 benefits GMI benefits 259,500,000 50 469,200,000 47 683,300,000 46 Total social 523,600,000 100 997,400,000 100 1,476,000,000 100 assistance benefits Notes: MDL rounded to nearest thousand, percent is rounded to nearest percent. To evaluate the effectiveness of the approximation to the government’s initial proposed two-month social support package to households, return migrants, and workers, and alternate durations of this package on mitigating the increase in poverty through the channels examined above, we start with the unmitigated effect under the “All” scenario, and combine it with the social assistance scenario for the duration of 2, 4, and 6 months. Under the 2, 4 and 6-month scenarios, 76,950, 73,550, and 74,650 families, respectively, benefitted from the GMI to varying degrees. The number of families benefitting under the 2, 4, and 6- month simulations is similar to the 75,700 families that received the GMI during April-May 2020 according to the Ministry of Labor and Social Protection. 24 The resulting national-level estimates suggest that with 2, 4, and 6 months of social assistance support there can be a tempering of the increase over baseline poverty to 1.6 and 0.5 and 0.2 percentage points in the 2, 4 and 6-month scenarios, respectively. In addition to mitigating the effect of the pandemic on the poverty rate, the social assistance measures also mitigate the increase in the depth of poverty, as measured by the poverty gap, from 1.2 percentage points under the unmitigated “All” scenario, to 0.7, 0.5 and 0.4 percentage points under the 2, 4 and 6-month scenarios, respectively. Figure 20: Simulated effects on the national Figure 21: Simulated effects on the national headcount of the “All” and the 2, 4, and 6- poverty gap of the “All” and the 2, 4, and 6- month social assistance packages relative to the month social assistance packages relative to the baseline scenario baseline scenario 3.0 1.4 2.7 1.2 Percentage point change over baseline Percentage point change over baselince 2.5 1.2 2.0 1.0 1.6 scenario 0.8 0.7 scenario 1.5 0.6 1.0 0.5 0.4 0.4 0.5 0.5 0.2 0.2 0.0 All All + SA All + SA All + SA 0.0 (2) (4) (6) All All + SA (2) All + SA (4) All + SA (6) Source: Authors’ estimates based on HBS 2018 data. The simulations reveal that the social assistance interventions are likely to have stronger impacts in rural areas, limiting the increase over baseline poverty to 1.5, 0.2 and 0.2 percentage points in the case of the 2, 4 and 6-month durations, respectively. The interventions are less effective in mitigating the effects on urban poverty, with increases of 1.6, 0.8 and 0.2 percentage points for the comparable durations. Similar patterns are also observed for the poverty gap -- in urban areas, the social assistance measures are less effective in mitigating the impact of the pandemic on the depth of poverty. 25 Figure 22: Simulated effects on urban/rural poverty headcount of the “All” and the 2, 4, and 6-month social assistance packages relative to the baseline scenario 5.0 4.5 Urban Rural Percentage point change over baseline 4.0 3.5 3.0 3.0 scenario 2.5 2.2 2.0 1.6 1.5 1.5 1.0 0.8 0.5 0.2 0.2 0.2 0.0 All All + SA (2) All + SA (4) All + SA (6) Source: Authors’ estimates based on HBS 2018 data. Figure 23: Simulated effects on urban/rural poverty gap of the “All” and the 2, 4, and 6-month social assistance packages relative to the baseline scenario 1.4 Urban 1.3 Rural Percentage point change over baseline scenario 1.2 1.2 1.0 0.9 0.8 0.7 0.6 0.6 0.5 0.4 0.3 0.2 0.2 0.0 All All + SA (2) All + SA (4) All + SA (6) Source: Authors’ estimates based on HBS 2018 data. The tempering in post-COVID simulated national poverty associated with the social assistance scenarios is likely due to two main factors. First, an increase in the number of individuals receiving unemployment benefits increasing the coverage of the unemployment benefit from 15,050 individuals in 4,050 households to 19,000 individuals in 5,450 households, 27,600 individuals in 7,850 households and 27,600 individuals in 7,850 households who were poor at baseline but not poor under the 2, 4, and 6-month scenarios. Second, the effect of the GMI, with 49,500, 50,200, and 45,500 individuals in 12,550, 13,800 and 12,950 households who were poor at baseline but not poor under the 2, 4, and 6-month scenarios receiving the benefit to varying degrees. Due to the small number of individuals who originally reported 26 receiving unemployment benefits it is likely that much of the reduction in poverty accruing to the unemployment benefit is due to increases in coverage rather than increases in generosity for existing beneficiaries. The stronger impact of the social assistance interventions on rural poverty can largely be attributed to poor rural households receiving greater per capita benefit amounts on average as compared with urban households. This impact is driven by more of the GMI benefits, which are relatively higher, going to poor rural households, while more of the unemployment benefits, which are relatively lower, are going to urban households. VII. How long will the return migrants stay and the welfare effects of their re-migrating? As containment measures are subsequently relaxed and borders reopened, return migrants may choose to depart or re-migrate for jobs in their previous host countries or seeking new destinations. As such, the re-migration of returnees and the accompanying resumption of remittances may help to offset some of the adverse effects of COVID-19. When surveyed about their intentions for the future, 67 percent of former migrants reported with some certainty that within the next 6 months they would return to their destination countries; an even higher share reported an intention of doing so within the next two years. Migrants who were abroad longer and who had higher levels of educational attainment were more likely to report an increased probability of returning (See Appendix Table A2). The majority of returnees planned to migrate back to their prior host country with 41 percent reporting it very likely and 31 percent reported quite possibly. Migrants also seemed optimistic about returning to the same job with 39 percent reporting that it was very likely and 44 percent reported it was quite possible. Figure 24: Proportion of migrants reporting likelihood of returning within the next 6 and 12 months 60 Percent of return migrants (%) 48 50 40 40 27 29 30 20 16 11 10 8 5 6 10 0 Very likely Quite Unlikely Not at all I can’t possibly answer Return next 6 months Return next 12 months Source: Staff computation using IOM migrant and remittance survey, 2020 For the majority of return migrants in the IOM return migrants survey, the decision to return back to their destination county does not appear to be influenced by the Covid-19 related economic crisis or fear that the migrants experienced in their destination countries. Focus group discussions with return migrants and members of their households reveal that the intentions to re-emigrate are driven by the socio-economic situation in Moldova, namely, small salaries, employment conditions, the quality of public services and 27 corruption. Among the respondents to the survey of return migrants, three quarters listed economic growth and improving living standards as factors that would influence their decision to stay in Moldova or return back to Moldova permanently, followed by improving employment opportunities and reductions in the level of corruption. More that a third similarly noted the need for improvements across a number of areas, including infrastructure, healthcare, education, social protection as well as safety / security. Figure 25: What would influence your decision to stay / return permanently to Moldova? Economic growth and raising living standards Creating employment opportunities Reducing corruption Improvements in healthcare field Increasing security / safety Strengthening the rule of law Improvements in the field of social protection Infrastructure investments Improvements in education field Others (specify) I can not answer 0% 10% 20% 30% 40% 50% 60% 70% 80% Share of return migrants Source: Data from the survey of return migrants, IOM / CBS Research. Migrants were also asked to estimate the monthly salary that would motivate a migrant to return and get a job in Moldova. The median reported salary is MDL 15,000 per month (average reported salary is MDL 16,500 per month), much higher than the average net salary in Moldova in 2020, according to data from the National Bureau of Statistics.26 There is some heterogeneity in these salary expectations – migrants from EU countries would expect a higher salary to consider returning, as compared to migrants from CIS countries. Similarly, migrants from urban areas, and especially from the capital city, have higher “reservation wages” compared to those from rural areas; these expectations also increase with the education level of the migrants, or with the average household expenditures of migrant’s household. 26 For details, see https://statistica.gov.md/newsview.php?l=ro&idc=168&id=6935 28 Figure 26: Average monthly salary that would motivate a migrant to return and get a job in Moldova 20,000 19,000 MDL per month 18,000 17,000 16,000 15,000 14,000 13,000 12,000 11,000 10,000 EU Balti Chisinau Other urban CIS Rural General Secondary Incomplete secondary Technical/vocational Tertiary Destination Region Education country Source: Data from the survey of households with migrants, IOM and CBS Research. We simulate the effect of return migration as a mitigating factor on the effects of COVID-19 by assuming based on the results of the 2020 IOM migrant and remittance survey, that 67 percent of migrants will return at some point in the year. Based on the simulations, the impact of migrants returning to their destination countries tempers the migrant scenario to an increase over baseline poverty of 0.2 percentage points and reducing the increase in poverty over baseline poverty to 2.1 percentage points under the “All” scenario, down from 2.7 points without considering re-migration. Similarly, the effects are tempered under the social assistance scenarios with changes of 0.9, 0.1 and -0.4 percentage points under the 2, 4, and 6-month scenarios, respectively. Similar patterns are evident when considering the effect on the poverty gap. 29 Figure 27: Simulated effects on national poverty Figure 28: Simulated effects on national poverty headcount across returnees re-migrating COVID- gap across returnees re-migrating COVID-19 19 scenario relative to the baseline scenario scenario relative to the baseline scenario Gap 2.5 Percentage point change over baseline 1.6 2.1 Change over baseline poverty gap 2.0 1.4 1.6 1.2 1.1 1.0 1.5 1.0 scenario 0.8 1.0 0.6 0.5 0.4 0.2 0.2 0.2 0.0 0.0 0.0 0.0 Remit Labor Migrant All Remit Labor Migrant All Source: Authors’ estimates based on HBS 2018 data. VIII. Experience of return migrants As some migrants may be unable to return to their destination countries in the immediate future, we consider the experiences of return migrants and their ability to integrate, focusing on migrants who have returned due to COVID-19 related reasons since March using the IOM migrant and remittance survey.27 To briefly recap, phone surveys using random digit dialing were administered to 2,004 migrant households between July and August 2020 and separately with 1,018 return migrants between August and September 2020. Only households that either had migrant members abroad or members who returned from abroad in the prior six months were enrolled. Similarly, the returnee survey was limited to migrants who returned over the same period. About a third of them returned due to reasons that can be attributed to the pandemic including job loss due to COVID, on temporary leave from job due to COVID and reduction in wages and working hours. The number of returnees were not as large as expected. Data from the Border Police of Moldova between March and September of 2020 indicated that the number of persons crossing over into Moldova was over 300,000 lower than the comparable period in 2019. The IOM survey also suggested a migrant return rate of 8 percent over the period March-August, implying that the pandemic, rather than inducing a surge of returns, instead put a pause on returns that happen as part of the natural course of temporary migration. A combination of travel restrictions and concerns about being able to return back to their host countries may have discourage large numbers of migrants from returning, even temporarily to visit their families on vacation. The return migrants surveyed were relatively well established in their host countries with 48 percent having been abroad for a period greater than 6 months (Appendix Figure A13). The most frequent duration 27 There was also a qualitative component with 8 focus groups and 20 in-depth interviews with returned migrants and family members with current migrants. A total of 65 people, 39 family members, 20 returned migrants and 6 migrants who left during COVID-19 were interviewed. 30 of migration was 6-12 months with 35 percent of migrants reporting this duration. The majority of these returnees were based in Russia (15%), Italy (18%), France (12%), Germany (12%), and the United Kingdom (11%) (See Appendix Figure A14). The main sectors of employment were male-dominated construction (43 percent) followed by domestic work (22%) which primarily employed women (Appendix Figure A15). Two-thirds of the returnees were male and the average age was 38.5 years. Most tended to come from less educated backgrounds—15 percent of the return migrants had primary education and 70 percent had some form of secondary education. When asked about their intentions during their stay in Moldova, 30 percent reported they would look for a job. Of those who reported that they would look for a job, just under two thirds reported that they would use acquaintances/relatives and a fifth reported they would use the media/internet. By contrast, only 7 percent reported that they would use the National Agency for Employment and less than 5 percent reported they would use private employment agencies. When considering the difficulties return migrants have encountered since returning, a quarter of them pointed to difficulty finding a job with a further 11 percent reporting the job they have found does not pay well and another quarter indicating financial difficulties more broadly (Appendix Figure A16). Despite some difficulties, 37 percent of migrants reported actually looking for a job and 61 percent of those who reported looking had done some paid work since returning home. Migrants also cited needing to pay for health insurance and negative attitudes from neighbors towards them or their family as one of the most frequent issues affecting them (Appendix Figure A18). Despite being entitled to unemployment allowance after fulfilling health insurance requirements only 10 percent reported having applied for it and of that, only 44 percent reported having received it. When asked the reasons for not applying for the benefit, a third stated that they were unaware of the benefit, 9 percent did not fulfil the health insurance policy requirement, 37 percent reported some form of distrust of state institutions and a quarter reported that they did not help from the state and a fifth reported not wanting to apply (See Appendix Figure A17). Focus group discussions with migrants reveal perceptions that the jobs that are available through the Employment Agency are not of good quality or well-paid, while ANOFM employees suggested, in in-depth interviews, that when good vacancies were available in IT, engineering and similar professions, there was a dearth of qualified candidates. In addition, there are spatial discrepancies in the demand for jobs and the supply of vacancies – according to data from ANOFM, some 6 out of 10 registered unemployed in September 2020 resided in rural area, whereas the majority of vacancies are in Chisinau, and the latter are not sufficiently well remunerated, once you account for additional costs of travel and/or lodging to motivate rural unemployed to move to Chisinau to take advantage of those vacancies. Just over 10 percent of return migrants report either having already invested money earned abroad in a business in Moldova, or intending to do so in the near future. Among these, half are operating, or intend to operate, in the production sector, primarily in agriculture. Among migrants who have or would like to open a business in the services sector, the main areas reported are transport services, followed by wholesale/retail trade. This reluctance to invest in a business in Moldova is driven by a number of factors, corruption being the most common among them, with more than half of the migrants who do not wish to invest in a business invoking it among top 3 reasons. In addition to corruption, a number of migrants invoke related concerns, such as not having anyone to defend their interests in front of the authorities, or high bureaucracy / reporting burden or too many controls. In addition, lack of own financial resources and difficulties in accessing bank loans, as well as high taxes are also common deterrents. 31 Figure 29: Three main reasons for not opening a business in Moldova. Corruption Lack of own financial resources Taxes are very high Too many controls Other (please specify) Bureaucracy/reporting burden Lack of knowledge, not prepared Nobody to defend my interests before authorities Unfair competition Difficulties in accessing bank loans I don’t know/I don’t answer 0% 10% 20% 30% 40% 50% 60% Source: Data from the return migrants survey, IOM and CBS Research. Note: Estimate based on the sample of all return migrants not having or intending to open a business. In line with the reasons preventing migrants from opening a business in Moldova, among some of the main policy actions that the migrants think the Government can undertake to help those who wish to return and open a business in Moldova, are reducing the tax burden, combatting corruption in the public administration, reducing the bureaucratic and reporting burden, and providing financial help, either by providing state subsidies, or by facilitating access to credit. In addition, some thirty percent of return migrants think that the government could help by encouraging fair competition, and an equal share report the need to de-monopolize some areas of economic activity. Finally, a quarter of respondents above cited insufficient knowledge to open a business, and, correspondingly, about a quarter of respondents state that the government could help by providing the necessary training courses on issues related to running a business. Figure 30: How can the state help citizens who wish to return and invest in businesses in Moldova? Reduce taxes Combat corruption in public administration Provide state subsidies Reduce the number of controls Reduce bureaucracy / reporting burden Facciliate access to credit Develop fair competition policies De-monopolize some areas of activity Provide training to those wishing to open a business Don't know / no asnwer Other (please specify) 0% 10% 20% 30% 40% 50% 60% Source: Data from the return migrants survey, IOM and CBS Research. Note: Estimate based on the sample of all return migrants 32 In this context, it should be noted that the government has a number of programs aimed at supporting investments by the diaspora and return migrants into the local economy, and aimed to help return migrants reintegrate into the labor market (see Box 1 for a summary of key initiatives). Among these, programs such as PARE 1+1 are specifically designed to stimulate the creation of new businesses in rural areas. However, awareness of these programs among return migrants appears to be limited. When asked whether they considered participating in entrepreneurial incentive programs such as PARE 1+1 and DAR, only 4 percent of return migrants answered in the affirmative, and almost two thirds stated that they were not aware that these programs existed. Box 1: Main government programs to support the diaspora and return migrant initiatives The Government of Moldova has a number of programs, both economic and cultural, aimed at engaging the return migrants and the diaspora. For instance, the program for diaspora children (DOR) supported summer camps for children from the diaspora, focused on cultural, linguistic and other activities, such as learning the national cuisine, meant to reconnect the children to Moldova. The Government has also organized at the local level, diaspora annual diaspora days, which included socio-cultural activities across a number of towns in Moldova, and also included round-table discussions with the diaspora o various issues. Likewise, Diaspora Excellence Groups were aimed to engage qualified diaspora members in offering consultative services to the Government on issues related to the design and implementation of public policies related to improving public services and promoting private sector development. Another program offered to Moldovan students studying at universities abroad internships for a duration of 1-6 months with government institutions in Moldova. In addition to these initiatives, several programs also aim directly at supporting business development or improvements of local service provision with the help of return migrants and the diaspora. The main programs, administered through the Bureau for Relations with the Diaspora, with support from development partners, include: DAR 1+3 (Diaspora acasa reuseste) program. A program about to be launched in 2020 aimed at local community development, investment of remittances in socio-economic development of rural areas in Moldova, and increasing the degree of participation of the diaspora in local development. The projects, to be proposed by local authorities and valuing at least MDL 300,000 would be co-financed by the diaspora (at least 10 percent of project budget), the local authorities (at least 10 percent of project budget), the government (at most 50 percent of project budget) and development partners (no fixed contribution). The eligible categories of projects will include: local infrastructure, environmental management, economy (regional development, agriculture), energy efficiency, culture, education or social protection. In evaluating the proposal, the committee will consider not only the relevance and the feasibility and sustainability of the project, but also, and with higher weights, the degree of the involvement of the diaspora in the initiation and implementation of the project, and the financial contribution of the diaspora to project financing. Diaspora Engagement Hub (DEH). DEH is a government program of thematic grants for citizens of Moldova residing abroad that has been implemented since 2017. Grants are offered in two categories: (i) professional return of the diaspora, aimed at financing the return of human capital from abroad for a period of 2 weeks – 2 months for the purposes of academic, social and economic development of Moldova; and (ii) innovative projects by the diaspora, based on the transfer of knowledge, experience and international good practices. The projects are similarly aimed to be implemented over the period of 2 weeks – 2 months. Grants in category (i) are up to MDL 40,000, and grants in category (ii) are up to MDL 50,000. 33 PARE 1+1 Program. The main flagship program targeting return migrants is the Government’s (PARE 1+1) program, supported by the International Organization on Migration (IOM) and running since 2010 under the slogan “Create your future at home”. It aims to help channel remittances toward stimulating the creation of new businesses in rural areas by mobilizing human and financial resources of Moldavian migrants in sustainable economic development of Moldova. The program provides a matching grant for each Leu invested by the migrants, up to the amount of MDL 200,000 (Euro 9,000). To date, the program has provided funding to more than one thousand businesses, with grants of over MDL 200 million, triggering further investments worth over MDL 700 million.28 Migration and Local Development Project (MiDL), implemented since 2015 supported by the UNDP and the SDC to assist communities affected by migration to benefit from improved essential local services (water and sanitation, health, social and education services), and have access to income- generating opportunities (employment registration and business development support). The project does so by providing support to the National Employment Agency at the national level and its territorial offices at the local level, and supporting local governments in a number of selected communities throughout Moldova in creating Hometown Associations that involve the diaspora in decisions about investments in improving local infrastructure and service delivery. During the first phase of the project, implemented during 2015-2018, 38 villages and cities throughout the Republic of Moldova took part in the project and established the first 38 Hometown Associations aimed at increasing the participation of the diaspora in local development and service provision. Another 101 local communities followed, as a spillover effect. The second phase of the project, to be implemented between 2019 and 2022, aims to expand the local component of the project to another 35 local communities. Several other programs have also been put in place by the Moldovan government, aimed at facilitating knowledge transfer and increased attachment to Moldova through the temporary return migration of the diaspora. One such program facilitates the temporary employment of Moldovan youth graduating from programs abroad in the public and the private sector in Moldova for a period of up to 6 months. The program is run through collaboration with various ministries and the NEA. The transport costs are covered for participants in the program, and they also receive a small stipend (European Commission, 2012). Another program, a joint collaboration between the IOM and the Academy of Sciences of Moldova, facilitates short-term visits for Moldovan scientists living abroad to Moldova to undertake research activities and knowledge transfer with their Moldovan colleagues (European Commission, 2012). IX. Conclusions The COVID-19 pandemic is simulated to have adverse impacts on the livelihoods of Moldovan households. Households who were already poor prior to the crisis are expected to be pushed deeper into poverty and those who would be classified as vulnerable prior to the crisis are at risk of falling into poverty. Increases in the headcount rate are more pronounced in rural areas largely due to the effect of the dual shocks of COVID-19 and the drought on rural employment and the disproportionate impact of remittances and 28 For details, see: http://eu4business.eu/news/1246-moldovan-businesses-funded-through-pare-11-programme- 2010 34 return migrants given preexisting regional tendencies.. If left unchecked, this could mean that close to an additional 100,000 people would fall into poverty. With the implementation of the proposed social protection package over the 2-month duration, the increase in poverty over the baseline rate can be tempered but not eliminated completely, pointing to need for prolonged social assistance support along with other complementary social services in order to prevent poverty from rising. The potential emergence of the “new poor” merits special policy attention given that their characteristics are likely to disqualify them from qualifying for existing social assistance programs that tend to be more geared towards the characteristics of the “traditional poor”. Specifically, based on our simulations, only 17, 16 and 9 percent of households that had at least one member suffer a job loss would qualify for the GMI benefits under the 2, 4, and 6-month social assistance scenarios, respectively. During this adjustment process, temporarily extending the duration and increasing the amount of unemployment benefits can help to support these individuals during the job search process particularly in the context of an unfavorable labor market. Adequate social security systems also need to be in place and maintained over the short-term to prevent the existing poor from falling deeper into poverty, particularly with respect to the COVID-19 shock compounded by the drought that affected agriculture productivity. As the economy recovers, since many of the individuals who were pushed into poverty on account of the pandemic suffered these welfare shocks mainly through labor market adjustments, active labor market policies with a focus on re and upskilling could help reintegrate into the labor market those who encounter difficulties with re-entry. Return migrants, particularly those who have been outside of the domestic labor market for an extended period, face a period of heightened uncertainty, confronted with both an unfamiliar and challenging labor market. The majority of return migrants expect to re-emigrate within a relatively short period of time, and very few are investing funds and knowledge acquired abroad into business opportunities in Moldova. While many of the migrants have tried to look for jobs in Moldova upon return, there is a clear mismatch in terms of both reservation wages and salaries being offered, as well as the spatial distribution of the unemployed and available vacancies, with many of the unemployed migrants from rural areas either unwilling or unable to take advantage of vacancies in Chisinau. Government programs, whether to mitigate the impacts of the pandemic through unemployment assistance, or to assist in starting a new business, or to find a job in Moldova, generally suffer from very low take-up. Many migrants rely on informal channels when they search for jobs and many are unaware of government incentive programs. Further support in the form of active labor market policies to migrants in order to assist them in finding jobs, as well as less stringent qualifying conditions to access the unemployment benefit while they search for jobs may improve the changes of return migrants of integrating into the domestic labor market and help support the welfare of their households. However, it should be recognized that increased social assistance support needs, which could be as high as US$4.25 million per month, representing a more than doubling of the current expenditure of US$2.1 million (World Bank, 2020) for a government with limited fiscal space. While supporting the efforts of return migrants to re-enter the Moldovan labor market is important, the welfare of those who remained abroad should not be overlooked. The resilience of remittances seen during this period are also a testimony to the hard work and sacrifices made by migrants abroad to continue supporting their families back home. Yet, the pandemic has left migrants vulnerable to joblessness, abuse or breach of contract by employers, as well as risks of contagion (World Bank, 2020c). 35 Aside from humanitarian considerations, as seen elsewhere - providing migrants access to adequate housing and health care is critical in keeping host communities safe from the pandemic. Some have been left stranded, unable to leave due to a combination of limited finances, travel restrictions or compromised legal status. Continued support needs to be extended to both those who remain and who wish to leave by the Moldovan government through its embassies and consulates abroad and where possible, in partnership with the host governments. Beyond the more immediate welfare impacts of the pandemic that were considered in this paper, the learning disruptions during the pandemic can have long-term impacts on human capital accumulation and can exacerbate pre-existing inequalities, as poor and vulnerable households tend to have a lower capacity to cope with school closures and other similar measures. According to nationally-representative data from the COVID-19 module of the HBS data for the 3rd quarter of 2020, some 10 percent of households reported that the household had limited access to distance learning because of internet connection deficiencies, and 8 percent reported not having sufficient electronic devices in the household to support distance learning needs. Importantly, lack of access is particularly acute among low-income households – households with a weak internet connection had per capita expenditures that were 22 percent lower than those who did not; households who were not able to take advantage of distance learning at all had per capita expenditures that were some 40 percent lower, on average (Table 3). These disparities are not only related to inequalities in access. More than 60 percent of households reported that distance learning educational materials were not understood by the students without the explanation of the teacher, and 29 percent of household reported that the students’ knowledge was insufficient to use the computer or online educational platform; low income households were more likely to report these difficulties related to student preparedness. Table 3: Difficulties with distance learning during the COVID-19 pandemic As the result of the COVID-19 pandemic, did your household Share of HH with school Expenditures per capita ratio experience the following difficulties with distance learning? children responding "Yes" between "Yes" and "No" respondents HH lacks access to distance learning due to lack of 4.5 -43.9 computer/tablet/phone/TV HH lacks access to distance learning due to lack of internet 4.5 -39.8 connection HH has limited access to distance learning due to insufficient 8.2 -32.8 number accessdevices of electronic HH has limited to distance learning due to weak 10.0 -22.0 Parental connection internet involvement in explaining educational materials was 53.3 -9.6 necessary Testing and homework verification by teachers was 24.8 1.4 sporadic/insufficient The student's knowledge was insufficient to use the computer 29.0 -9.8 or materials platform online educational Educational are not understood by the student 61.0 -8.4 without the teacher's explanation rd Source: Estimates based on the COVID_19 module in the 3 quarter HBS 2020 data. These differential effects on learning can perpetuate inequality traps and have a negative effect on inter- generational mobility. A recent study based on data from Latin American countries estimates that , intergenerational persistence in education – a measure of the extent to which the educational attainment of individuals is independent of the education of their parents -- could increase by 7 percent (Neidhoefer et al. 2021). In Moldova, and in the ECA region more broadly, intergenerational persistence in education has been increasing for the most recent cohorts, in contrast to the trends over the same period in middle- and high-income countries (Narayan et al., 2018). The pandemic could exacerbate these trends, given pre- 36 existing disparities and the more pronounced impacts of COVID-19 on low-income households, and reinforces the need to invest alleviating disparities in human capital. Figure 31: Intergenerational mobility (relative) by age cohort, 1940-1980. Notes: Relative mobility is measured by the extent to which the educational attainment of individuals is independent of the education of their parents, using the coefficient from regressions of children’s years of education on the education of their parents. Higher values of this regression coefficient indicate greater persistence, and hence lower relative mobility. For details, see Narayan et al. (2018). Source: Authors’ estimates based on the Global Database of Intergenerational Mobility (GDIM). References Bussolo, Maurizio; Koettl, Johannes; Sinnott, Emily. 2015. Golden Aging: Prospects for Healthy, Active, and Prosperous Aging in Europe and Central Asia. Washington, DC: World Bank. © World Bank. Cantarji, Vasile, Natalia Vladicescu, Maria Vremis and Veaceslav Batrinescu. 2020. “Impact of the COVID- 19 pandemic on migration: mobility, number and profile of returned migrants, specific vulnerabilities of groups affected by the decline in remittances. CBS Research, Chisinau, Moldova. Cojocaru, Alexandru, Mikhail Matytsin, 2018. "Moldova Poverty and Shared Prosperity Update 2018," World Bank Other Operational Studies 30468, The World Bank. European Commission. 2012. Social Impact of Emigration and Rural-Urban Migration in Central and Eastern Europe: Moldova Final Country Report. Frankel, J. 2009 “Are bilateral remittances countercyclical? NBER Working Paper No. 15419, Cambridge, MA. International Labor Organization. 2017. Migrant Workers: The Case of Moldova. 37 International Organization for Migration. 2017. “Extended migration profile of the Republic of Moldova 2009-2014.” International Organization for Migration. 2020. “IOM rapid field assessment of the impact of COVID-19 on the wellbeing of the Moldovan diaspora: an evidence base regarding migrants’ coping strategies and contributions.” Chisinau, Moldova. Ministerl Afacerilor Interne al Republicii Moldova. 2017. Compendiul Statistics al Profilului Migrational al Republicii Moldova pentru anii 2014-2016. Narayan, A., Van der Weide, R., Cojocaru, A., Lakner, C., Redaelli, S., Mahler, D. G., Ramasubbaiah, R., Gupta N., Thewissen, S. 2018. Fair Progress?: Economic Mobility Across Generations Around the World. Equity and Development. Washington, DC: World Bank. Neidhoefer, G., N. Lustig, and M. Tommasi. 2021. “Intergenerational transmission of lockdown consequences: Prognosis of the longer-run persistence of COVID-19 in Latin America.” ECINEQ Working Paper No. 571. World Bank. 2010. Harnessing the Diaspora for Development in Europe and Central Asia. World Bank. 2016a. Moldova Poverty Assessment 2016: Poverty Reduction and Shared Prosperity in Moldova: Progress and Prospects. World Bank. 2016b. Moldova – Path to Sustained Prosperity: A Systematic Country Diagnostic. World Bank. 2017. A Human Rights-Based Approach to the Economic Security of Older People in Moldova. World Bank. 2020a. Poverty and distributional impacts of COVID-19: Potential channels of impact and mitigating policies. Mimeo. Poverty and Equity Global Practice. World Bank. 2020b. Moldova Macro-Poverty Update, Annual Meetings. Washington, DC: The World Bank. 38 Appendix: Additional figures and tables Figure A1: Incidence of Covid-19 in the population Figure A2: Employment by select industries in Moldova, 2019 Agriculture, forestry and fishery 9 Industry 21 Construction 23 Trade, hotels and restaurants 15 Transportation and Communication 7 Public administration, 7 Education, Health and social work 19 Other Source: National Bureau of Statistics 39 Figure A3a: Simulated effects on national poverty Figure A3b: Simulated effects on national poverty headcount across alternate remittances and gap across alternate remittances and other other COVID-19 scenarios relative to the baseline COVID-19 scenarios relative to the baseline scenario scenario 3.5 Gap Percentage point change over baseline scenario 3.0 2.9 1.6 1.4 1.3 Change over baseline poverty gap 2.5 1.2 1.0 2.0 1.0 1.6 1.5 0.8 0.6 1.0 0.7 0.4 0.4 0.2 0.5 0.2 0.1 0.0 0.0 Remit Labor Migrant All Remit Labor Migrant All Source: Authors’ estiamates based on HBS 2018 data. Fig A4: Distribution of per capita consumption as a percent of the poverty line under the baseline (2018) 20 18 16 14 Percent of individuals 12 10 8 6 4 2 0 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Consumption relative to the poverty line Source: HBS 2018 data. 40 Fig A5: Distribution of employment status of Fig A6: Distribution of employment by sector of new poor vs existing poor new poor vs existing poor 15 18 24 34 3 47 6 Existing 75 Existing 73 poor poor 82 New 18 2 New poor poor Employed Unemployed Outside Labor Force Agriculture Industry Construction Services Source: Author’s estimates based on HBS 2018 data. Fig A7: Age distribution of new poor vs existing Fig A8: Distribution of educational attainment of poor new poor vs existing poor 5 16 25 24 27 27 33 44 Existing Existing poor poor 70 57 New 42 29 New poor poor Child Working Age Elderly Primary Secondary Tertiary Source: Author’s estimates based on HBS data. 41 Fig A9: Distribution of educational attainment Fig A10: Distribution of employment types of within sectors new poor vs existing poor 120 100 30 80 41 59 35 66 60 76 40 54 1 62 Existing 20 34 41 3 69 24 poor 0 4 New Agriculture Industry Construction Services poor Paid employee Primary Secondary Tertiary Non-paid employee/fam Self-employed Source: Author’s estimates based on HBS 2018 data. Fig A11: Distribution of employment types within Fig A12: Distribution of employment in sectors sectors within regions 7 7 2 59 27 35 31 43 85 14 93 92 14 36 14 24 32 22 41 35 26 23 21 12 Agriculture Industry Construction Services Agriculture Industry Construction Services Paid employee Self-employed Other Nord Centru Sud Chisinau Source: Author’s estimates based on HBS 2018 data. 42 Fig A13: Distribution of return migrants by duration of stay abroad 0 3 10 25 35 27 1-3 months 3-6 months 6-12 months 1-3 years 3-5 years 5-10 years Source: Author’s estimates based on IOM migrant and remittance survey. Fig A14: Distribution of return migrants by destination country 20 18 16 14 12 10 8 6 4 2 0 Source: Author’s estimates based on IOM migrant and remittance survey. 43 Fig A15: Distribution of return migrants employment by destination country Education Other I don’t know / I don’t answer Unemployed Agriculture Transport Trade Services Industry Domestic care sector Construction 0 10 20 30 40 50 Source: Author’s estimates based on IOM migrant and remittance survey. Fig A16: Problems experienced by return migrants I don’t know / I don’t answer 0 Other 2 I could not get medical help 3 Problems/ conflicts with authorities 3 Unable to benefit unemployment allowance/social… 4 The job I have is low paying 11 Bad attitude from neighbors towards me or my family 13 I haven’t experienced any problems 20 The need to pay for health insurance 20 Lack of financial resources 24 Difficulty to find work in Moldova 25 Impossibility of travel for 14 days because of quarantine 66 Source: Author’s estimates based on IOM migrant and remittance survey. 44 Fig A17: Reasons for not applying for unemployment insurance benefit as reported by return migrants Source: Author’s estimates based on IOM migrant and remittance survey. Table A1: Aggregation of sectors into Agriculture, Industry, Construction and Services Sector Agriculture Industry Construction Services Agriculture, hunting, forestry and finishing ✓ Manufacturing ✓ Construction ✓ Transportation ✓ Wholesale and retail trade ✓ Accommodation and food services ✓ Real estate activities ✓ Profession, scientific and technical activities ✓ Administrative and support service activities ✓ Art, entertainment and recreation ✓ Other service activities ✓ Activities of households as employers ✓ Mining and quarrying ✓ Utilities ✓ Information and communication ✓ Finance and insurance activities ✓ Public administration ✓ Education ✓ Human health and social work ✓ Activities of extraterritorial organizations ✓ 45 Table A2: Probit of the likelihood a migrant will return to work abroad Migrant Variable will return Female 0.0284 (0.2856) Age 0.0671 (0.0647) Age Squared -0.0007 (0.0008) Upper Secondary 0.3502 (0.2657) Tertiary 0.7531** (0.3509) 6 months-1 year 0.6974*** (0.2230) Over 1 year 0.6496** (0.3122) Construction 0.0488 (0.3719) Other -0.2756 (0.3322) CIS/Other 0.1542 (0.2869) Constant -1.5286 (1.4083) Source: Author’s estimates based on HBS 2018 data. Note: *, ** and *** corresponds to significance at the 10%, 5% and 1% level, respectively. Robust standard errors are in parentheses 46