Bangladesh and Covid-19: Reversals and Resilience*

Atiya Rahman,1 Nusrat Jahan,2 Imran Matin3 and Hossain Zillur Rahman4


This article examines two primary data sets to identify the effect of the Covid-19 pandemic on different sectors and vulnerable populations in Bangladesh. It attempts to identify how the trends in sectors such as agriculture, ready-made garments (RMGs), education, employment among youth, and women’s participation have changed due to the pandemic compared to pre-Covid-19 levels. The results show that the agriculture and RMG sectors demonstrated resilience due to sustained government policies. In contrast, the other sectors, such as education, youth employment, and women’s participation in the labour market, have been negatively affected, leaving a long‑term consequence for the country’s development. The article concludes with suggestions for inclusive and targeted policies, and community-based approaches to pre-empt new challenges to make development progress in Bangladesh.


Covid-19, new poor, livelihoods, youth, Bangladesh, women, Covid-19 recovery.

1 Introduction

Before the outbreak of the Covid-19 pandemic, Bangladesh had experienced macroeconomic stability for over a decade – sustained high growth, stable single-digit inflation, small budget deficits, and an improved balance of payments. As a result, Bangladesh was one of the fastest-growing economies globally and the second-fastest-growing economy in South Asia (World Bank 2019). Human development indicators also improved significantly – there was a decline in extreme poverty, a sharp increase in life expectancy, and an increase in women’s education and economic participation. However, the pandemic has brought significant disruption in this development trajectory, slowing down growth and reversing the poverty reduction trend for the first time in two decades.

Since the onset of the pandemic, the BRAC Institute of Governance and Development (BIGD), BRAC University, and the Power and Participation Research Centre (PPRC) have been jointly conducting a telephonic panel survey on nationally representative rural and urban slum households to capture the changing impact of the pandemic on low-income communities in Bangladesh.5 This article utilises the three-round telephonic survey on 6,038 households, of which 2,649 are from rural areas and 3,389 are from urban slums. It also utilises the survey conducted by BIGD in February–March 2021, funded by the Covid Collective, with 1,929 young men and women aged 18–25, to understand the impact of the pandemic on youth employment, with an emphasis on female youth.

PPRC-BIGD survey findings show that the agriculture sector – which was performing well before the pandemic, partially because of strong, sustained government policies and support – showed resilience against the economic shock induced by the pandemic. However, the pandemic revealed the fragility of vulnerable groups such as the urban poor, women, and young people, for whom support systems have been either missing or fragmented. A significant share of the population who were marginally above the poverty line before the pandemic fell below the line during the pandemic and are yet to recover, indicating the emergence of a possible poverty trap.

Urban slum dwellers are more affected than rural citizens because of the nature of jobs and higher non-food expenditures, notably rent, in cities. Reverse migration is another consequence of the pandemic. Urban slum dwellers are changing locations, possibly to cut expenditures. The youth are also experiencing steeper challenges in recovering their incomes, employment, and skill‑building opportunities. Finally, we find that women are facing the brunt of the pandemic: almost a third of young women employed before the pandemic were out of a job as of January 2021, according to a separate youth survey carried out by BIGD on a nationally representative youth sample.

Another major area of concern is human capital accumulation. School closure for almost 18 months, since March 2020, may have long-lasting consequences on children – including learning loss and possible income loss in their adulthood. It may also increase inequality as access to digital learning tools is sensitive to economic status. We also find an indication of nutrition deficiency among urban slum dwellers, many of whom have reduced food expenditure. Children of poorer households may also experience long-term loss of mental and physical capabilities, which may impair their earning potentials in adulthood. The challenge thus is not only about addressing reversals but also building forward pathways more effectively.

The article argues that Bangladesh, with large gaps in its existing system, must build on the unique strengths of that system, notably a strong grass-roots presence of non-governmental organisations (NGOs), and develop a statecraft that is decentralised, innovative, and in a constant learning mode. The article aims to explore the areas where possible reversals are taking place and identify which socioeconomic groups were hit the hardest, which groups exhibit the resilience to cope, and what the potential sources of resilience are.

The article is structured as follows. In Section 2, the surveys used in this article are discussed. In Section 3, the evidence is presented on the resilient agriculture segment of the Bangladesh economy. Next, in Section 4, the major trends in reversal and emerging traps are explained, while in Section 5, the impact of the pandemic is described from a demographic perspective. Section 6 concludes with final remarks indicating possible pathways to mitigate the challenges.

2 The surveys

2.1 Livelihoods survey

Figure 1 shows a snapshot of the Covid-19 scenario in Bangladesh over time and when the surveys – used in this article – happened. The figure illustrates the relevance of these surveys in the context of the pandemic. This article primarily uses PPRC‑BIGD’s nationally representative three-round panel survey on the livelihoods of Bangladesh’s urban slum and rural households in the Covid-19 context. These telephonic surveys started off as a methodological innovation responding to the constraints of in‑person surveys under Covid-19 restrictions.

See the full explanation in the long description

The first round of the livelihood survey conducted in April 2020 included a sample of 12,000 households, equally representing urban slums and rural areas. The urban sample was randomly drawn from BIGD’s census from October 2016 to January 2017 with 24,283 households in 35 slums – where BRAC’s Urban Development Programme (UDP) had operations – across nine districts in Dhaka, Chattogram, Khulna, Barishal, and Rangpur divisions. The rural sample was drawn from BIGD’s nationally representative survey of 26,925 rural households across 64 districts of all eight divisions, conducted between October 2017 and January 2018. We randomly drew 6,000 households from the rural database. Of the total sample, 5,471 households were interviewed over the phone.

For the second round, in June 2020, an additional 6,000 households were drawn from the same data sets – 4,000 from urban areas and 2,000 from rural areas. A larger urban sample was selected to facilitate disaggregated analysis of the urban centres. Of the total sample of 11,471 households – 5,471 previously interviewed and 6,000 new households – 7,545 were interviewed.

The third round of the survey was conducted 11–31 March 2021. The 7,545 households that were interviewed in the second round were resurveyed to see the trend in their circumstances. Out of those resurveyed households, 6,038 households were interviewed, of which 3,549 households were surveyed in all three rounds, while the remaining 2,489 were surveyed in the second and third rounds.

In every round, three standard modules were included in the survey instrument – livelihoods (employment, income, expenditure), coping mechanism, and migration. In the third round, an additional module was included on the educational life of children.

The household head was the respondent in the survey, where 16 per cent of the households are female-headed, which is slightly higher than the national statistics (13 per cent). Since this survey interviewed the household head, an analysis of this data set on groups such as women and youth would not provide a complete picture. For this, we used another nationally representative survey, funded by the Covid Collective, on the young population. The sampling technique of BIGD’s youth study is described next.

2.2 Youth survey

The sample for the youth survey comes from BIGD’s youth survey in 2018, which adopted a multi-stage random sampling technique. We collected information on youth from five regions of Bangladesh, including the north (Rangpur and Rajshahi divisions), south (Khulna and Barishal divisions), northeast (Sylhet division), southeast (Chattogram division), and central (Dhaka and Mymensingh divisions). Thirty sub-districts were randomly selected from each region. One union – the lowest administrative unit – was then randomly selected from each selected sub-district, and one village/mahalla (neighbourhood) was randomly selected from each union. In each village/mahalla, 14 youth aged between 15–35 years6 were randomly selected. Thus, the survey interviewed a total of 4,200 youth, equally distributed across the five regions.

For the youth survey during the pandemic, we utilised the phone numbers of this database and conducted a telephonic survey on those who were 18–35 years as of 2021. BIGD conducted this follow-up survey on 3,795 youths between 27 February and 7 March 2021.

2.3 Survey mode

The pandemic has created an urgency for gathering high-frequency data to understand the evolving economic and social changes, and ensuring rapid policy engagement. Since in-person surveys are usually expensive and time-consuming, telephonic surveys can help reach the population and contribute to rapid analysis (Rahman et al. 2021). Though telephonic surveys have unique challenges, standard protocols can overcome those hurdles.

Up-to-date telephone numbers are crucial to achieve a standard success rate. Thus, the most recent sample frames can be advantageous. Moreover, we tried to reach out to the samples three times in a row on different days during the survey period to improve the success rate. We created a short, clear, and concise questionnaire and ensured the active engagement of the pre‑testing team to finalise it.

Lastly, to protect privacy, we masked the phone numbers by using a technical programme, SurveyCTO, and stored the data in an encrypted server. Following our data protection protocol, we collected the mobile number, name, age, and other biological information.7 We have a strong protocol against sharing any data containing personally identifiable information with any external party.

These approaches contribute to collecting reliable, high-quality data, which is crucial to reflect on the rapid changes in this crisis period.

3 Resilience in the agriculture sector

Since the relaxation of the nationwide lockdown in June 2020, the economic recovery in Bangladesh has been sluggish and unevenly distributed among all the occupations (Rahman et al. 2020b). A change in income from June 2020 (hereafter referred to as the early recovery stage) to March 2021 (hereafter referred to as the late recovery stage), compared to the pre-Covid-19 level, February 2020, shows that agriculture is the only sector that experienced positive growth a year into the pandemic (Figure 2). In other occupational categories, especially among transport workers, skilled labourers, small business owners, and housemaids, income is yet to recover to the pre-Covid-19 level.

Percentage change in income in the order by category in June 2020 and March 2021. Agriculture: -33, +12; Transport work: -48, -15; Skilled labour: -43, -22; Unskilled labour: -50, -11; Factory work: -18, -1; Salaried jobs: -16, -5; Small businesses: -49, -16; Rickshaw pullers: -48, -19; Housemaids: -61, -12.

The agriculture sector in Bangladesh is expected to be less affected by the pandemic as maintaining social distance is less likely to be a constraint given the nature of the activities. There are several other propositions to explain the resilience of this sector – for example, low connectivity between local agriculture and the global economy, continued inflow of remittances (Malek, Truong and Sonobe 2021), existing social safety net programmes (SSNPs), and the lower infection rate in rural areas during the first wave.8 However, the agriculture sector did experience delayed harvest due to labour shortages, difficulties in selling products and procuring inputs due to disruptions in transportation, and higher costs. Farm families also saw reductions in remittance and non-farm business sales (ibid.), but the government immediately took special measures to ensure enough labour supply for harvesting, transportation for inputs, and marketing (Uddin, Shoaib and Arafat 2020).

Food security is a politically sensitive issue in Bangladesh since the devastating famine in 1974 (Ahmed, Haggblade and Chowdhury 2000; Hossain 2017). The Government of Bangladesh (GoB) has been adopting policies for long-term advancement in this sector through smooth management of input supplies, input subsidies, fair output prices, improving credits, promotion of food and non‑food crops, and so forth. As a result, the country achieved self-sufficiency in rice production by the late 1990s (Rahman et al. 2014). In 1999, Bangladesh came up with its first comprehensive agriculture policy, with a clear goal of ensuring food security as well as firm incomes. Additionally, a quiet agricultural revolution has taken place in response to natural calamities, sociopolitical changes, population growth, new technology and new opportunities in the rural non-agricultural sector, commercialisation, and changes in macro policy and sector policy reforms (Ministry of Agriculture 2006). Consequently, rice production has increased threefold between 1971 and 2018 (BBS 2019b).

The agriculture sector is the lifeblood of Bangladesh’s rural economy. Thus, its resilience has strengthened rural economic recovery. In the late recovery phase, we find that rural households experienced 12 per cent higher income than the pre-Covid-19 level, while the income level was still 14 per cent lower than the pre-Covid-19 level in urban slums.

4 Reversals in socioeconomic parameters

In this section, we discuss the pandemic-induced reversals in the following areas:

4.1 Reversal in poverty alleviation

The national statistics show that both extreme and moderate poverty rates in Bangladesh have been decreasing since 1991. The reduction in poverty rate only became sluggish between 2010 and 20169 (Panel C of Figure 3). The slowdown is due to the slower reduction of the urban poverty rate (Panels A and B of Figure 3).

The PPRC-BIGD survey shows a two-percentage-point rise in poverty a year into the pandemic. In our urban slum sample, the increase in poverty is eight percentage points, while poverty decreased in our rural sample during the same period.

The reason behind this slow reduction of urban poverty is that rural poor people have migrated from villages to urban slums, resulting in higher population density in the latter. Push factors (i.e. poverty and unemployment in villages) have been more influential than pull factors (i.e. better job opportunities and facilities in cities) in encouraging rural–urban migration (Farhana, Rahman and Rahman 2012), leading to an increase in urban poverty.

Moreover, global evidence shows that the urban poor tend to have less access to national SSNPs, especially cash transfers, compared to their rural counterparts (Gentilini, Khosla and Almenfi 2021). In Bangladesh, the SSNPs are rural-centric; only 11 per cent of urban households received SSNP support, while the rate was 35 per cent among rural households in 2016 (BBS 2019a). Besides, there are inclusion errors; approximately 14.5 per cent of SSNP beneficiaries receiving the assistance are not eligible for the support (BBS 2011).

Our survey also looks at the poverty dynamics across four income groups. First, we construct our extreme and moderate poverty groups based on the BBS’s lower and upper poverty lines. Second, we define those who were above the upper poverty line and below the median income level as the vulnerable non-poor. Finally, we define the non-poor as those who were above the median income level. Figure 4 shows that the overall proportions of vulnerable non-poor and non-poor households reduced drastically in June 2020, right after the first lockdown, with income falling below the extreme poverty level for two thirds of the population. By January 2020, most of the households were able to recover their income substantially, and the poverty rate dropped. Yet, it was two percentage points below the level in February 2020.

4.2 Emergence of the ‘new poor’

Our data points out a possible reversal in the poverty trend. During the early recovery stage, all income groups experienced a drastic drop in their income. However, during the late recovery stage, the scenario did not improve for the pre-Covid-19 vulnerable non‑poor. A significant segment of this group remains in the poverty trap, especially in urban areas. Figure 5 shows that 59 per cent of urban slum dwellers who were vulnerable non-poor before the pandemic were now below the poverty line; the rate is 44 per cent in rural areas. We identify this group as the ‘new poor’, i.e. those who were the vulnerable non-poor prior to the pandemic. The results show a slow recovery of the vulnerable non‑poor only because this group is at greater risk of falling back into longer-term poverty.

We calculate a national estimation for the ‘new poor’ based on the transition of the pre-Covid-19 vulnerable non-poor into poverty. The estimated national poverty rate was 20.5 per cent in 2019. Thus, 29.5 per cent of the population are estimated to be in the vulnerable non-poor income group – between the upper poverty line and the median income level – before the pandemic. We find that 50 per cent of the pre-Covid-19 vulnerable non‑poor households in our survey fell below the poverty line in March 2021. Thus, the estimate of the national ‘new poor’ stands at 14.75 per cent,10 i.e. 24.5 million individuals.

Our survey further finds that the households mainly relied on savings, household income, and informal support from neighbours and relatives to meet their food demand during the lockdown period. A negligible number of people reported assistance from the government to meet their food demand. The insufficient amount of assistance and increase in debt has been increasing the vulnerability of the poor group.

4.3 Reverse migration

Between 2000 and 2019, the share of the urban population in Bangladesh increased from 23.6 per cent to 37.4 per cent (World Bank 2022), but this trend seems to have reversed as the Covid-19 virus emerged.

Many people in the big cities moved back to less expensive cities because of the economic distress induced by the pandemic. Our study finds that around 27 per cent of the urban slum dwellers migrated over the past year. Out of the total, 20.5 per cent migrated before June 2020, and an additional 6.8 per cent have followed since. These groups are referred to as early and late migrants, respectively, throughout this article. Of the total number of migrating urban households, 17.5 per cent have already returned, resulting in a net reverse migration of 9.8 per cent from urban areas over the year.

Forty-three per cent of the early migrants moved to villages, whereas 60 per cent of the late migrants have done the same. This may indicate that the early migrants hoped to find job opportunities in other cities, where the costs were also lower. Indeed, most of the early migrants moved from large metropolitan cities such as Dhaka and Chattogram to lower-cost district towns (Rahman et al. 2020b). However, the late migrants adjusted their strategy and moved to their native villages, where their living costs are likely to be the lowest.

The census data on the urban slum dwellers show that about one fifth of the total household expenditure was towards rent and utilities before the pandemic (Figure 6). The non-food expenditure of the urban household has doubled during the late recovery period compared to the early recovery period. The drastic increase in costs is ubiquitous across all categories: rent, utilities, and health costs have gone up by 46 per cent, 51 per cent, and 81 per cent, respectively. This sharp increase is perhaps due to deferred rent payments during the early months of the pandemic. Indeed, the second round of the PPRC-BIGD survey conducted in June 2020 found that 26 per cent of the urban households were not paying rent at that time, indicating that their rent payments were piling up. A similar incidence happens with the utility costs and education costs. We find that the cost of education for urban slum dwellers increased by 33 per cent in March 2021 compared to the pre-Covid-19 level.11

4.4 Reversal in human capital accumulation

The schools were closed in Bangladesh from March 2020 until September 2021, and were shut down again in February 2022. The extended closure has had far-reaching consequences, increased risks of learning loss and dropout, as well as growing psychological issues and economic costs. The survey on children’s education focused on a subset of the larger sample of 6,099 households in the third round of the study – 4,940 households in rural areas and urban slums who had children of school‑going age.

Before the pandemic, 14 per cent of the primary and 21 per cent of the secondary school-going aged children in the surveyed households were not in school, indicating deficiencies in human capital from the outset. The deficit was more pronounced in urban slums and among children of secondary school age. When Covid-19 hit the world, the school closures affected students in ways that are likely to pose a higher risk of long-term learning loss; namely, not studying at all, pursuing only unsupervised self‑study, or studying irregularly in any other mode of study. Thus, in mentioning these scenarios, we define children as being at risk of learning loss. Our findings show that at least 19 per cent of primary and 25 per cent of secondary students face learning loss risk during school closures.

Based on the national statistics of primary and secondary school-going children in 2019 (BANBEIS 2020), our calculation predicts that nationally, approximately 3.42 million primary and 2.50 million secondary students, a total of 5.92 million school-going children, are at this critical risk. The actual learning loss risk is likely to be higher because we only consider the ‘unsupervised’ self-study component here. The risk is more distinct in urban slums and among secondary school-going children.

Some children completely stopped formal education, whereas other children resorted to alternative mediums of education, including unsupervised self-studying. Besides these two mediums of studying – not studying at all and unsupervised self-study – we find a few more alternative modes of studying during school closures. We categorise them into four groups: studying with family member support (parents, siblings), distance learning (TV, online, internet), coaching/private lessons, and shifting to madrasas (Islamic learning institutions that remained open).

Family support to compensate for school closures is higher in primary school students (28 per cent) than secondary school students (7 per cent). Mothers are found to play a significant role in supporting primary school children to study at home. Parental education might be an influential factor in helping children with their studies; however, we have not taken this dimension into consideration for this study.

We further find a low level of compliance with distance learning, including TV, online, and the internet. Overall, only 10 per cent of students reported being engaged in distance learning to compensate for school closure, even though the government started broadcasting distant classes on the national television channel early into the pandemic. Moreover, most of those who reported distance learning mentioned the online classes, which mostly dominates among the high socioeconomic groups.

Market mechanism (coaching/private) has been the prominent alternative to learning at school; 51 per cent of primary and 61 per cent of secondary school children took private classes/coaching. Urban slum children (61 per cent) were found less likely to use the market mechanism than rural children (71 per cent).

Finally, only 4 per cent of primary school students and 1 per cent of secondary school students shifted to madrasas during the school closure over one year of the pandemic.

In our youth survey, less than 5 per cent of the youth (aged 15–35) mentioned taking any kind of training since the pandemic started, even though about a quarter of the sample were students who were facing the year-long school closure. This data also indicates the dismal scenario of human capital accumulation in Bangladesh during the pandemic.

5 The predicament of the female population and youth

National statistics show that 11 per cent of the youth were unemployed in 2016–17, and about 36 per cent of working-age women participate in the labour force compared to 81 per cent of working-age men, and the unemployment rate is much higher than that of men (BBS 2018). We utilise BIGD’s youth survey, conducted between February and March 2021 during the late recovery stage, to understand the impact of the pandemic on the youth and female populations.

5.1 Impact on youth

Among the respondents who said they were working before the pandemic, 48 per cent remained in income-generating activities (IGAs) during the first lockdown, and even seven months after the economy resumed, 85 per cent of them got back to IGAs12 (Figure 7).

This population includes mostly youths working in the education sector – teachers and private tutors. As the educational institutes had been closed since March 2020, we found a high unemployment rate among the teachers and the private tutors. Additionally, many youths making handicrafts or working in private firms and factories remained out of work, possibly because many firms in these sectors produce non-essential products or services (Figure 8). Overall, youths in sectors that involve skilled workers (education, private firms, and crafts) were more affected and faced persistent job loss and underemployment (lower working hours), compared to sectors employing less skilled people (day labourers, agriculture labourers, and transport workers).

The third round of the PPRC-BIGD survey, conducted in 2021, found that per capita income was 7 per cent below pre-Covid levels. We analysed youth’s individual income from all occupations in the three time periods. We found that income fell sharply during the lockdown in April 2020. Hence, we compare the average change in individual youth income in two time periods – pre-pandemic in February 2020 and seven months after the resumption of economic activities in January 2021. For the two time periods, we compared the change in average monthly income between two groups: (i) youths in IGAs in February 2020 (could be still working or not working in January 2021) and (ii) youths in IGAs in both time periods.

Youths who were working before the pandemic faced an 18 per cent decrease in the average monthly income between February 2020 and January 2021 (Figure 8). The rate is 2.5 times higher than the general sample in the PPRC-BIGD survey. We found that 15 per cent of youths working before the pandemic were no longer in IGAs. Hence, this figure reflects the fall in the average income of those youths, as well as those who were working in January 2021.

In contrast, the percentage decrease in the average income of youths who were working in both time periods was 11 per cent. Although lower than the decrease in the full sample, the data implies a substantial income loss even among youths who were working in both time periods, even seven months after economic activities resumed. We also find that youths living in urban areas faced a greater decrease in average income between the two periods.

Almost a third of the employed population in Bangladesh have no formal education, and only 5 per cent have tertiary education (BBS 2018). The high school dropout rate (38 per cent) in the country is still very high (Daily Star 2018), and less than 5 per cent and less than 2 per cent of people aged 15 and above have tertiary and vocational training, respectively (BBS 2018).

Bangladesh has allocated about 12 per cent of its total budget to education, which stands at just 2 per cent of gross domestic product (GDP) in 2021–22. Globally, the average government expenditure is 4.5 per cent GDP, and the spending in Bangladesh is one of the lowest in the world (World Bank 2021). Even amidst the pandemic, the budget allocation for education remains unchanged in terms of size and percentage of GDP.

The country ranks 116th among 183 countries on the index of quality education for the youth in the Youth Development Index (The Commonwealth 2021). In the Global Competitiveness Ranking, Bangladesh scores poorly in the categories of competitive skills (ibid.).

5.2 Loss in employment and income of female youth

Women’s participation is concentrated in certain sectors; 92 per cent of the working women in Bangladesh are employed in the informal sector, whereas the rate is 85 per cent for working men – meaning only 8 per cent of working women are in the formal sector, less than half the rate of working men (BBS 2018).

In Bangladesh, the ready-made garments (RMG) industry is the main source of formal employment for women – 58.30 per cent of RMG workers are women (New Age 2020). With the cancellation of RMG orders when Covid-19 hit, women in the RMG sector lost their jobs. The International Labour Organization (ILO) predicts that many low-skilled women may never get back their job (ILO 2020).

A gender-disaggregated analysis shows a disproportionate impact of the pandemic on young women compared to young men. Forty-seven per cent of the male youths working before the pandemic were still working in April 2020, while the rate is 52 per cent for the female youth (Figure 9). Though a larger percentage of working female youth could retain their livelihoods during the lockdown, the difference was minor.

However, by January 2021, more male youths managed to recover their livelihoods compared to female youths. Of youths who were involved in IGAs before the pandemic, 89 per cent of the young men and 71 per cent of the young women had got back to IGAs by January 2021 (Figure 9). Almost a third of the female youth working prior to the pandemic did not find work seven months after the extended lockdown was lifted. This rate is almost three times as high as the male youth working before the pandemic.

Consequently, we see a disproportionate negative impact of the pandemic on young women’s income. Figure 10 illustrates the percentage decrease in income for both groups. Among those who were in IGAs in February 2020, young women’s average income decreased by 38 per cent, more than twice as much as their male counterparts who saw a decrease of 17 per cent.

In the case of female youths, jobseekers in both periods – the percentage decrease in average income was more than twice as high as their male counterparts (Figure 10).

Both our result and existing literature depicts a fragile labour market scenario for women. According to BBS (2018), the estimated unemployment rate for young men was 10 per cent, which is much lower than the rate among young women (14 per cent). In the BIGD survey, we also find that only 19 per cent of the young women were engaged in IGAs pre-pandemic, whereas 56 per cent were engaged in housework.

The gendered nature of work across industries may partially explain why the livelihoods recovery among young women is low. According to UN Women (2020), women are overrepresented in many of the hardest-hit industries during the pandemic and this fact is consistent with our survey results where we identify that occupations like private tutoring, teaching, handicraft making, and factory work have greater numbers of female participants and were hit the hardest due to the pandemic. The pandemic has made it more difficult for women to retain employment, leading to the risk of permanent unemployment because of economic and non-economic barriers to getting back to work.

6 Conclusion

This article discusses four major domains of development – agriculture, social protection, job creation, and education. Although agriculture, like other industries, was negatively affected by the pandemic initially, the sector recovered quickly because of well-established systems and supportive policies developed over the decades. The RMG sector, too, has followed a similar trend during the course of the Covid-19 pandemic.

The resilience of the agriculture and RMG sectors in the face of the pandemic can be explained by the politics of policy priorities in Bangladesh. Historically, food security has been a priority for all governments. The impetus to avoid crises such as the famine of 1974 has ensured policies that support long-term growth and advancement in the sector. Agriculture is the lifeblood of the rural economy, which in itself has an influential voice in national-level discourse, and the presence of systems and mechanisms that can respond swiftly to disasters has resulted in a quick turnaround despite the pandemic.

The RMG sector, too, has become critical to Bangladesh’s economic success in the last two decades. The sector’s substantial contribution to GDP, employment, and female labour participation, along with the considerable influence of business owners in matters of policy, has made RMG a priority for governments for a number of years now. As a result, the RMG sector has received timely support from the government, in the form of stimulus packages.

However, this has not been the case for other domains. In Bangladesh, SSNPs are rural-centric. There is no national system to include the urban poor and protection schemes explicitly designed for informal sector workers and vulnerable non‑poor, a demographic that constitutes a significant portion of the population. The vulnerable non-poor, in particular, living dangerously close to poverty levels and being almost as likely as the poor to be hit hard by prolonged crisis, flew below the policy radar.

With the high unemployment fuelled by the pandemic, women and the youth are experiencing more challenges to get back into the labour market. While digital platforms have become essential for survival during the pandemic, it is more difficult for women to access digital devices and labour market participation due to the existing gender divide. A youth survey finds that 24 per cent of young women use the internet, while the rate is 56 per cent among young men (Matin et al. 2019).

The larger negative impacts of the pandemic highlight the importance of comprehensive social assistance programmes. NGOs have been the custodians of development in Bangladesh, especially in times of crisis. In contrast, state response has often been constrained by limitations of systemic corruption and a lack of compliance and enforcement. Moreover, state responses are often too centralised to reach target populations who are vulnerable to crises. The current pandemic points to the limitations of the existing weak and fragmented system that have become more apparent now. Similarly, the NGO-driven model of disaster response and social support appears to be no longer adequate to effectively address the multifaceted challenges of a global crisis.

Due to the increasing integration into a global economy with its own vulnerabilities, and as the epicentre of climate change, Bangladesh is likely to face more frequent shocks of this magnitude in the future. It must adopt a multi-actor, community-based, integrated model of development that embraces embedded informalities and is appropriate for the challenges of modern times.

The pandemic has shown that a horizontal approach that utilises the strengths of all available actors – from community to NGO – can be successful in delivering life-saving assistance to the most vulnerable. For a rapidly escalating and unpredictable crisis such as the pandemic an adaptive response strategy is needed, which allows for old and new networks of organisations to align and work collectively with minimum loss of lives (Khan et al. 2021).

Strong grass-roots-level presence and activity are needed to provide urgent, mass support during a national crisis. Innovations to address some of the practical, operational challenges of delivery, such as targeting, can be a good point of entry. There is a need to empower local governments, independent agencies, and ministries. Local states must have the operational freedom to customise interventions based on local needs. Adaptive governance processes and decentralised authority can cope with contingent and local needs better.

The prolonged school closure during the pandemic has worsened the education sector and further increased the inequality in learning. This is why learning beyond the classroom is crucial to recover the learning loss. A local community-based learning environment can be an effective solution. Children belonging to a strong peer network find more value in their education and are less likely to drop out. Interventions should emphasise the communal aspect of learning, and promote cooperation and help-seeking behaviour.

Moreover, a hybrid model of both in-person and remote learning is a critical need, as a significant number of students do not have access to digital devices. Educational content and teaching materials need to be improved and interactive. The challenge of dropouts will become prominent as schools reopen. The dropout rate may increase because of students’ inability to cope with either the learning or the financial crisis. Thus, catch-up education is going to be crucial. These catch-up classes should be scheduled out of school hours. Additionally, fiscal interventions are a necessary tool to reduce dropouts. Widening student stipends and increasing the amount of the existing stipend will help poor households continue sending their children to school.

What approaches will help vulnerable households fight against unprecedented shocks is an intriguing research question that deserves further exploration. Institutional reforms depend to an extent on the state’s capacity to assess risk and vulnerability, a synergy with and among the local communities and other development actors, and the existence of efficient community-based feedback mechanisms. Although the comparative benefits of a community-based, horizontal model of development in Bangladesh have manifested during the pandemic, the mechanisms involved in such a complex, multi-actor system are not understood well yet. There is a scope for further research that explores how to best leverage the strengths of all available actors by adopting a community-driven approach to tackling crisis and adversity.


* This IDS Bulletin was funded by the UK government’s Foreign, Commonwealth & Development Office (FCDO) through the Covid Collective. The Collective brings together the expertise of UK- and Southern-based research partner organisations and offers a rapid social science research response to inform decision-making on some of the most pressing Covid-19-related development challenges. The Covid Collective cannot be held responsible for errors, omissions, or any consequences arising from the use of information contained. Any views and opinions expressed do not necessarily reflect those of FCDO, the Covid Collective, or any other contributing organisation. For further information, please contact:

1 Atiya Rahman, Associate Research Fellow, BRAC Institute of Governance and Development (BIGD), BRAC University, Bangladesh.

2 Nusrat Jahan, Head of Communications and Knowledge Management, BIGD, BRAC University, Bangladesh.

3 Imran Matin, Executive Director, BIGD, BRAC University, Bangladesh.

4 Hossain Zillur Rahman, Executive Chairman, Power and Participation Research Centre (PPRC), and Chairperson, BRAC, Bangladesh.

5 For details on sampling, see Rahman et al. (2020a) and Rahman et al. (2020b).

6 According to the National Youth Policy, ‘any citizen between 18 and 35 years of age shall be regarded as Youth’. See National Youth Policy 2017: Draft English Version.

7 The personally identifiable information includes any external, biological, physical, or any other information that can individually or jointly identify the person.

8 The eminent economist, Hossain Zillur Rahman, highlighted these five points as the rationale for the resilience of this sector in an interview (Wasif 2021).

9 The latest Household Income and Expenditure Survey (HIES) was conducted in 2016. The next survey, HIES 2020–21, is currently under preparation.

10 Fifty per cent of the size of the pre-Covid vulnerable non-poor (29.5 per cent), equal to 14.75 per cent, which we refer to as the ‘new poor’.

11 We inflate the cost of education collected in 2017 (baseline survey) and consider the inflation-adjusted cost as the cost at the pre-Covid-19 level.

12 This percentage is based on the sample of youths who were in IGAs in February 2020, i.e. 41.5 per cent of the entire sample.


Ahmed, R.; Haggblade, S. and Chowdhury, T.-E. (eds) (2000) Out of the Shadow of Famine: Evolving Food Markets and Food Policy in Bangladesh, Baltimore MD: Johns Hopkins University Press

BANBEIS (2020) Bangladesh Education Statistics 2019, Dhaka: Bangladesh Bureau of Educational Information and Statistics, Ministry of Education, Government of Bangladesh

BBS (2019a) Final Report of the Household Income and Expenditure Survey 2016, Dhaka: Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of Bangladesh (accessed 13 April 2022)

BBS (2019b) Yearbook of Agricultural Statistics 2018, Dhaka: Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of Bangladesh (accessed 13 April 2022)

BBS (2018) Labor Force Survey Bangladesh 2016/2017, Dhaka: Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of Bangladesh

BBS (2011) Report of the Household Income and Expenditure Survey 2010, Dhaka: Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of Bangladesh (accessed 13 April 2022)

BBS (2007) Report of the Household Income and Expenditure Survey 2005, Dhaka: Bangladesh Bureau of Statistics, Statistics and Informatics Division, Ministry of Planning, Government of Bangladesh

Daily Star (2018) ‘42pc Girls Drop Out Before SSC’, 28 February (accessed 13 April 2022)

Farhana, K.M.; Rahman, S.A. and Rahman, M. (2012) ‘Factors of Migration in Urban Bangladesh: An Empirical Study of Poor Migrants in Rajshahi City’, Bangladesh e-Journal of Sociology 9.1: 105–17 (accessed 13 April 2022)

Gentilini, U.; Khosla, S. and Almenfi, M. (2021) Cash in the City: Emerging Lessons from Implementing Cash Transfers in Urban Africa, Social Protection & Jobs Discussion Paper 2101, Washington DC: World Bank Group (accessed 13 April 2022)

Hossain, N. (2017) The Aid Lab: Understanding Bangladesh’s Unexpected Success, Oxford: Oxford University Press

ILO (2020) ‘ILO Monitor: Covid-19 and the World of Work, Sixth Edition’, Briefing Note, Geneva: International Labour Organization (accessed 13 April 2022)

Khan, M.; Roy, P.; Matin, I.; Rabbani, M. and Chowdhury, R. (2021) ‘An Adaptive Governance and Health System Response for the Covid-19 Emergency’, World Development 137: 105213 (accessed 13 April 2022)

Malek, M.A.; Truong, H.T. and Sonobe, T. (2021) Changes in the Rural Economy in Bangladesh under Covid-19 Lockdown Measures: Evidence from a Phone Survey of Mahbub Hossain Sample Households, ADBI Working Paper 1235, Tokyo: Asian Development Bank Institute

Matin, I. et al. (2019) Youth of Bangladesh: Agents of Change? Youth Survey 2018, Dhaka: BRAC Institute of Governance and Development (accessed 13 April 2022)

Ministry of Agriculture (2006) Actionable Policy Brief and Resource Implications: Agriculture Sector Review, Dhaka: Ministry of Agriculture, Government of Bangladesh

New Age (2020) ‘Number of Female Workers in RMG Sector Falling: Surveys’, 15 December (accessed 13 April 2022)

Rahman, H.Z.; Matin, I.; Banks, N. and Hulme, D. (2021) ‘Finding Out Fast About the Impact of Covid-19: The Need for Policy-Relevant Methodological Innovation’, World Development 140: 105380 (accessed 14 April 2022)

Rahman, H.Z. et al. (2020a) Livelihoods, Coping and Support During Covid-19 Crisis (Phase I), Dhaka: BRAC Institute of Governance and Development (accessed 13 April 2022)

Rahman, H.Z. et al. (2020b) Livelihoods, Coping and Recovery During Covid-19 Crisis (Phase II), Dhaka: BRAC Institute of Governance and Development (accessed 13 April 2022)

Rahman, M.M.; Klees, B.; Sal-sabil, T. and Khan, N.A. (2014) ‘Rice, Smallholder Farms, and Climate Change in Bangladesh: Thala of Policy Options’, Journal of Indian Research 2.1: 59–66

The Commonwealth (2021) Global Youth Development Report 2020, London: The Commonwealth Secretariat (accessed 19 April 2022)

Uddin, J.; Shoaib, M. and Arafat, M.F. (2020) Impacts of Covid-19 on Agriculture in Bangladesh, Dhaka: World Overview of Conservation Approaches and Technologies (accessed 14 April 2022)

UN Women (2020) ‘Covid-19 and its Economic Toll on Women: The Story Behind the Numbers’, 16 September (accessed 14 April 2022)

Wasif, F. (2021) ‘2020 Sale shahorer cheye gram keno bhalo’ [Why were villages better than cities in 2020?], Prothom Alo, 18 July (accessed 14 April 2022)

World Bank (2022) Urban Population (% of Total Population) – Bangladesh (accessed 27 April 2022)

World Bank (2021) Government Expenditure on Education, Total (% of GDP) (accessed 14 April 2022)

World Bank (2019) South Asia Economic Focus, Fall 2019: Making (De)centralization Work, Washington DC: World Bank (accessed 14 April 2022)


© 2022 The Authors. IDS Bulletin © Institute of Development Studies. © Crown Copyright 2022. | DOI: 10.19088/1968-2022.125

This is an Open Access article distributed for non-commercial purposes under the terms of the Open Government Licence 3.0, which permits use, copying, publication, distribution and adaptation, provided the original authors and source are credited and the work is not used for commercial purposes.

The IDS Bulletin is published by Institute of Development Studies, Library Road, Brighton BN1 9RE, UK This article is part of IDS Bulletin Vol. 53 No. 3 July 2022 ‘Pandemic Perspectives: Why Different Voices and Views Matter’.