A snapshot of labour market trends throughout the pandemic

The Mauritian economy has realised an unprecedented structural transformation since independence. Today, the agricultural sector accounts for about 3% of GDP, the industrial sector contributes about 20%, while the remaining 77% of GDP is produced by the services sector. Trade, financial and insurance activities and tourism services together account for almost 40% of the country’s GDP, followed by manufacturing, public administration, health and education services, real estate activities and administrative and support services activities. With the outbreak of COVID-19 and border closures, there were virtually no tourist arrivals between April and October 2020.^{Footnote 5} Hospitality and food services, tour operators, taxis and shops selling to tourists have been heavily affected. In comparison with 2019, accommodation and food services activities contracted by 65.8% in 2020. Manufacturing exports, wholesale trade and construction activity are expected to partially rebound in 2021.^{Footnote 6}

Before the pandemic struck, labour force participation had increased over time, reaching 59.3% in 2019 while unemployment had declined from 7.3% in 2009 to 6.7% in 2019. Gender gaps had narrowed over the past decade, with women’s labour force participation increasing from 42.9% to 46.2% between 2011 and 2019. In 2019, over 80% of the employed population worked for a wage, about 15% were self-employed, while the remaining 5% were distributed between employers (about 3%) and contributing family workers (about 2%). The services sector, led by trade, accommodation and food services activities and in third position transport, contributed over 70% of total employment. The secondary sector followed with 24% of total employment, with manufacturing and construction constituting the lion’s share, while agriculture contributed about 6% in 2019.^{Footnote 7}

The outbreak of the pandemic and the subsequent economic crisis have taken a toll on employment. The number of employed declined by 13.9%, approximately 74,000 people, from the first quarter of 2020 to the first quarter of 2021 (Table 1).

Table 1. Employment, unemployment and inactivity, from first quarter 2020 to first quarter 2021.

Source: Based on data from the first quarter of 2020, 2021 of the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

Activity rate: labour force over population; Unemployment rate: unemployed over labour force.

In the first quarter of 2020, the employed population aged 15–64 not in full-time education was estimated at 534,802 persons. In May, when the country was still in lockdown, total employment declined to 405,387 (−24%). Although employment bounced back to 523,725 employed persons in December 2020, the resurgence of cases and the re-introduction of pandemic-related measures led to a severe drop in employment to 460,729 (−12%) during the first quarter of 2021. The loss in employment has been accompanied both by an increase in the number of unemployed and in the number of inactive individuals. Whereas the unemployment rate increased from 7.2% in the first quarter of 2020 to 10.8% during the first quarter of 2021, the inactivity rate dropped from 73.4% to 63% during the same time span, indicating that displaced workers stopped looking for new work and left the labour force in greater numbers.

Analysing labour demand during this time, measured as the number of advertised vacancies, several patterns emerge. In line with the observed employment outcomes (Figure 1), job postings sharply contracted from a monthly average of over 3000 over the period October 2019 to February 2020 to less than 800 between March and May 2020. Although labour demand picked up in June (2125) and August 2020 (2292), generally, demand continued to trend downward from September 2020 (824) into the first quarter of 2021 (quarter mean: 617).

Figure 1. Trends in number of advertised vacancies, Q4/2019–Q1/2021.

Source: Based on data from Employment Service Monthly Bulletins of the Ministry of Labour, Human Resource Development and Training of the Republic of Mauritius.

The number of vacancies represents the aggregate number of public and press vacancy announcements from the last quarter of 2019 to the first quarter of 2021.

Numbers of vacancies for Q4-2019, March–May 2020 and Q1-2021 are a monthly average over the period.

Women too, have suffered negative labour market effects as a consequence of the pandemic. The percentage decline in employment has been similar for women and men during the initial phase of the pandemic. Between the first quarter and July 2020, about 6.5% of women and 7.2% of men lost their jobs. Although the labour market was seemingly recovering in the latter half of the year, a total of 559,636 individuals reported losing their employment between September 2020 and the first quarter of 2021. Analysing these specific displacement patterns, several observations stand out. First, in absolute numbers, the total amount of jobs lost due to the pandemic and the ensuing economic crisis was higher among men given that the pre-crisis employment rate was higher for men. However, while most men who had lost their jobs started to look for new work, and thus remained engaged in the labour market, most women exited the labour market and became inactive after displacement.

The number of unemployed men rose by 83% compared with a growth of about 6% among women until July 2020. Even though these growth effects levelled off in the second part of the year, this trend was reversed at the beginning of 2021. Between December 2020 and the first quarter of 2021, a total of 41,830 fewer men and 21,166 fewer women reported being employed (Table 2). Given that unemployment rates remained stable in comparison to December 2020, increased job losses translated into a drop in the activity rates for both men (−15%) and women (−10%). Although the female activity rate was estimated at 58.3% in July 2020 and recovered to 61.4% – outpacing its pre-pandemic level (61.0% in the first quarter 2020) – female activity rate saw a sharp decline in the first quarter of 2021 reaching 51.5%.

Table 2. Employment, unemployment and inactivity, by sex and age group, first quarter 2020 to first quarter 2021.

Source: Based on data from the first quarter of 2020, 2021 of the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

Activity rate: labour force over population; Unemployment rate: unemployed over labour force.

At a disadvantage in terms of access to jobs before the pandemic, Mauritian youth aged 16–24 years were not spared by the crisis. Youth employment declined by about 33% between the first quarter of 2020 and the first quarter of 2021, compared with a drop of 10% among workers in other age groups. In 2020, the activity rate among youth declined substantially as a result of COVID lockdown measure until May and gradually bounced back to 80.7% in December. The number of youth looking for a job increased by about 756, and the unemployment rate rose from 24.1% in May to 33.1% in December. Again, although the recovery throughout 2020 positively impacted the Mauritian labour market, these positive trends have been entirely reversed during the first quarter of 2021. Youth were not spared from these developments, their employment decreased by 18,726 between December 2020 and the first quarter of 2021 leading to an increase in youth unemployment from 33.1% to 48.9% during the same time. As unemployment and inactivity heavily influence future outcomes at a young age and youth inactivity can permanently impair productive potential and influence lifetime patterns of employment, pay and job tenure, the sharp decline in youth labour force participation to 37.9% warrants further policy attention.

In terms of sectoral distribution, the reduction in employment during the first 3 months since the start of the pandemic was largely attributable to the dynamics of informal employment (Statistics Mauritius and World Bank, 2020). Whereas formal employment dropped by about 10% from 384,000 in the first quarter of 2020 to 343,817 in May 2020, informal employment sharply contracted from 150,800 to 61,570 during the same time (−59%). By September 2020, formal employment had recovered and in the first quarter of 2021 was above pre-crisis levels. By contrast, informal employment increased between June and December and then contracted again and fell to 53,184 in the first quarter of 2021 (Figure 2).

Figure 2. Trends in employment levels by formality status, Q1/2010–Q1/2021.

Source: Based on data from the first quarter of 2020 and of 2021 of the Continuous Multi-Purpose Household Survey and six rounds of the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

The determinants of labour market outcomes during the pandemic

To substantiate our initial findings, we analysed the correlates of labour force participation and corresponding employment and displacement patterns through a series of regressions.^{Footnote 8} To structure our analysis, we performed regressions on four selected time windows. Starting with the widest possible time frame of all available survey months in 2020, we separately analysed the first quarter of 2020 (i.e. Pre-pandemic time window); the first period after the initial lockdown (i.e. May until July 2020); the second period after the lockdown had been (partially) lifted in August (i.e. September until December 2020); and the first quarter of 2021. This way of analysing the data allowed us to study the responsiveness of labour market outcomes and derive insights into the effectiveness of the implemented policy measures to the ongoing pandemic (e.g. the labour market effect of lockdowns). Given existing sex differences in the Mauritian labour market, we also provide the results of our analysis for subgroups of men and women. Due to data limitations, our analysis of displacement patterns is restricted to the time window between September and the first quarter of 2021.

We report our findings in Figure 3 and the full results in the Supplemental Table A2. In line with prior findings, men were more likely to participate in the labour force. For instance, men across all groups were 12.6% more likely than women to be in the labour force at the outset of the pandemic (p < 0.01). According to our estimation results, slightly more women than men left the labour force during the pandemic, with this effect most pronounced in the aftermath of the first lockdown (i.e. between May and July 2020). This can be attributed to women’s greater prevalence of displacement and household-related factors. Interestingly, these sex differences in labour force participation level off for divorced, separated and single women during the pandemic.

Figure 3. Correlates of labour force participation by sex and time period.

Source: Based on data from the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

The point estimates are based on OLS regression including month and district fixed effects. The dependent variable in all model specifications is a respondent’s answer to the question of whether they have lost their employment due to business closures between September 2020 and the first quarter of 2021. At the outset of the pandemic no significant differences existed between women based on marital status, however by the September–December period women that were either divorced, separated and/or single were between 18.9% and 21.1% more likely to be active participants in the labour market than their married counterparts. We cannot detect such a pattern among men. With respect to education, higher education seems to be a robust predictor of labour force participation, even when considering the pandemic. For instance, in comparison to their peers without a completed primary education and/or no schooling, women with post-secondary or tertiary education were almost 30% more likely to be active participants in the labour force (see Supplemental Table A2). Although this advantage varies across educational attainment it varies little during the course of the pandemic, and in comparison to men, the effect is approximately double in size. The observed education advantage for men ranged between 8.9% and 21.7% between the first quarter of 2020 and corresponding first quarter in 2021. As expected, those aged 25–44 years were more likely to be active in the labour force than their peers aged 16–24 and 45–64, even though this effect is subject to moderation once the pandemic progressed. Also, these age cohort effects seem not to differ between men and women. In line with this finding, across all model specifications, household size seems to be a robust predictor of labour force participation. Again, no significant differences existed between women and men.

With respect to our analysis on employment, we report our findings in Figure 4 and the full results in Supplemental Table A3. Although men were more likely to participate in the labour force, these differences did not translate into more employment opportunities after the onset of the pandemic. While men were 3.6% more likely to be employed during the first quarter of 2020, this effect disappeared in the direct aftermath of the first lockdown between May and July 2020. Interestingly, the gains in employment between September and December 2020 led to the re-emergence of existing gender biases with respect to employment. Although at the pandemic’s outset divorced, separated and single women were less likely to be employed in comparison to their married peers, they were just as likely to be employed as the pandemic progressed. A notable exception were single women, who were 16.7% more likely to be employed in comparison to their married peers, a remarkable difference compared to the first quarter of 2020, during which single women were 14.8% less likely to be employed. In sharp contrast, we can detect a somewhat opposite pattern among single men who were 7.1% less likely to be employed than their married peers between September and December 2020 but showed little difference during the first quarter of 2020.

Figure 4. Correlates of employment by sex and time period.

Source: Based on data from the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

The point estimates are based on OLS regression including month and district fixed effects. The dependent variable in all model specifications is a respondent’s answer to the question of whether they have lost their employment due to business closures between September 2020 and the first quarter of 2021.With respect to education, higher education was a more robust predictor of employment, especially during the pandemic. Interestingly, the education effect on employment was equally pronounced between men and women across all levels of education whereas these differences did not seem to matter at the outset of the pandemic. For instance, in comparison to their peers without a completed primary education and/or no schooling, women with post-secondary or tertiary education were almost 10% more likely to be employed between September and December 2020. Although this advantage varied across level of educational attainment, a lower or upper secondary education was more conducive for employment during the pandemic. At the outset of the pandemic, in the first quarter of 2020, those aged 25–44 and 45–64 were more likely to be employed than youth. However, this effect was subject to modest moderation once the pandemic progressed. These findings are in line with recent research on youth employment dynamics in Mauritius (Reference Castaneda, Gaddis and RanzaniCastaneda et al., 2020). In addition, there were no differences in these age cohort effects between men and women. In line with this finding, across all model specifications, household size was not a robust predictor of employment and again, no significant differences existed between women and men.

The observed drops in employment during the pandemic warrant further investigation. For this reason, we analyse specific displacement patterns in this section. Due to data limitations, this analysis is limited to between September 2020 and March 2021.

Again, our findings confirm the notion of an asymmetric displacement pattern between women and men (Figure 5). For example, we find that women were 12.1% more likely to lose their job when compared to their male peers. Importantly, women with less education were most susceptible to displacement during the pandemic. For example, women who had attained an upper secondary education or higher were up to 10% less likely to be displaced compared to women without any or an incomplete primary education. Furthermore, we find that married women were more likely to lose their employment compared to their divorced (−24.3%), separated (−24.1%) and single peers − (−23.3%).

Figure 5. Correlates of displacement by sex, May–March 2021.

Source: Based on data from the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

The point estimates are based on OLS regression including month and district fixed effects. The dependent variable in all model specifications is a respondent’s answer to the question of whether they have lost their employment due to business closures between September 2020 and the first quarter of 2021.

Displacement was most likely to occur for those aged 45–64 years (11.0%), with this effect more pronounced for men (11.5%) and less so for women (8.7%). These findings are notable to the extent that re-employment and re-training for this age cohort might prove to be more challenging (Reference Bednarzik, Kern and HisnanickBednarzik et al., 2021). When looking at the sectoral composition of displacement, the effect was most pronounced in the services sector. As expected, most displaced workers were employees and again, women were disproportionally represented among those workers who were displaced (i.e. approximately 65% of displaced employees). Synthesising these insights, displacement appears to be concentrated in the services sector and among women, with women who are the least educated the most vulnerable.

Further analysis of displacement patterns between September 2020 and the first quarter of 2021 revealed that women were more likely to be displaced than men, indicating the presence of a potential gender bias in firms’ layoff decisions. As expected, men and women with higher education were significantly less likely to be laid off during this period, underscoring the importance of education for increased job security. Interestingly, we find that older employees were better protected from being laid-off in comparison to their younger peers, reinforcing the importance of tenure/work experience for increased job security. This effect holds for both men and women. The upshot in unemployment corresponds to intensifying displacement in the secondary (or manufacturing) sector. According to estimation results, being employed in the secondary sector increased the likelihood of being displaced by 9%.

Against this background, a key insight of our analysis is the stark differences between men and women in their displacement and employment. Although initial differences in gender patterns can be attributed to gender roles, where women are the primary providers of childcare and other unpaid work, the pandemic has proven to be a powerful catalyst in undermining any advances in levelling the labour market playing field.^{Footnote 9}

Protecting jobs and living standards in the short and medium-term

Following the outbreak of COVID-19 in Mauritius, the Government implemented several measures to contain the spread of the virus, including adopting a sanitary curfew; the cessation of public transport; international border closures; the closure of schools, universities and shopping malls; suspension of onsite work among public and private sector workers except for essential activities; and mass testing.^{Footnote 10}

To contain the socioeconomic impact of the pandemic, the government intervened promptly with fiscal support measures.^{Footnote 11} These included a moratorium on capital and interest payments on existing and new loans; a loan facility benefiting micro, small and medium enterprises; the distribution of food packages to the most vulnerable households; and compensation schemes for workers. This section focuses on the compensation schemes, namely, the GWAS and the SEAS, which were introduced to protect jobs and support incomes.^{Footnote 12}

Temporary wage subsidies are an adequate instrument for providing support for worker incomes during crises, such as the one triggered by the COVID-19 pandemic, that affect both labour supply – in this case, the ability of workers to go to work – and labour demand – because firms are obliged to close. Under the GWAS, private sector firms are entitled to cash transfers to pay their wage bills. In particular, private-sector employees with basic salaries up to MUR 25,000 were entitled to receive their full salaries, employees taking in basic monthly salaries between MUR 25,000 and MUR 50,000 received a fixed amount of MUR 25,000, while employees with basic salaries above MUR 50,000 a month were not eligible to receive wage subsidies from the government. Under the SEAS, the self-employed aged 18 years and over (i.e. including those aged 18) and in business during the 3 months prior to the outbreak were entitled to receive monthly financial support of MUR 5100 as long as they were not: beneficiaries of government assistance, dependent spouses, belonging to a household with a monthly income above MUR 50,000, fishermen, full-time university students or formal private-sector employees.

Based on pre-crisis labour force survey data, virtually all workers employed in low-skill jobs and most workers performing medium-skill jobs received a monthly basic salary of up to MUR 25,000 and were therefore eligible to receive full subsidies under the GWAS.

Since July 2020, the schemes have only covered workers in the tourism sector. During the second lockdown in March 2021, the support was provided to other sectors as a reduced payment at half the original monthly payment to wage workers, as well as a one-off grant of MUR 10,000 to the self-employed. With the start of the reopening in April 2021, a full month assistance was provided under both GWAS and SEAS for all sectors and the schemes were extended until September 2021 for tourism-related companies.

Between 20% and 50% of workers in high-skill jobs typically received a basic monthly salary between MUR 25,000 and MUR 50,000 and were therefore eligible to partial salary replacement. These workers were in households from the middle and top of the household income distribution. The GWAS, therefore, appears to be well-targeted toward the workers most in need. Workers in high-skill jobs were also more likely to be able to work from home as about 79% used computers at work, compared to only 30% of workers in medium-skilled jobs and less than 6% in low-skilled jobs (based on the Continuous Multi-Purpose Household Survey, first quarter, 2020).

According to data from the Mauritius Revenue Authority published by Statistics Mauritius, about 270,000 private sector workers benefited from the GWAS in March 2020, about 256,000 in April and about 220,000 in May. Based on 2019 labour force survey data, this corresponds to 84% coverage of private-sector employees with a basic salary below MUR 50,000 (in March), 80% (in April) and 69% (in May) (Statistics Mauritius, 2020).^{Footnote 13}

The SEAS provided minimum income replacement to both formal and informal self-employed workers. The monthly transfers of MUR 5100 corresponded to about 50% of the minimum wage; this is considerably less than the median (MUR 14,630) and average monthly (MUR 18,390) business income (in 2020 prices) reported by the self-employed in 2019. Nonetheless, providing the self-employed with a cash transfer in a time of crisis is critical to ensuring these people receive income, as the self-employed cannot operate their businesses during lockdowns and face considerable uncertainty over both their ability to reopen and the availability of customers while demand remains weak.

About 203,000 SEAS applications were processed in March, over 186,000 in April and about 185,000 in May. These numbers are considerably higher than the 97,900 employers and own-account workers who were estimated to have been operating in Mauritius in 2019 (Statistics Mauritius, 2020) and still considerably higher than the 108,200 if contributing family workers were considered as potential applicants. This gap may be attributable to informal employees applying for income support under SEAS rather than the GWAS. Informal employees are private-sector workers whose employers do not pay contributions to the National Pension Fund. Such workers are not registered with the Mauritius Revenue Authority and are therefore not eligible to benefit from the wage subsidies provided through the GWAS. Instead, informal employees are likely to have applied for financial assistance under the SEAS.^{Footnote 14} Based on the 2019 Continuous Multi-Purpose Household Survey data, about 91,000 private sector workers can be categorised as informal according to the criterion explained above. Adding these informal employees to the number of self-employed leads to a total number of potential applicants that is closer to the number of applications processed under the SEAS.

It is extremely important that financial assistance schemes reach all workers, particularly those at the bottom of the income distribution, including informal wage workers. However, it is worth noting that the amount provided under the SEAS to informal private sector employees is equivalent to 50% of the minimum wage and therefore only a fraction of what these workers usually make per month and of what was provided to formal employees.

In October 2020, the government announced several funding initiatives to be implemented over the following 8 months to reduce unemployment. As a result, from 1 November 2020, until 30 June 2021 five initiatives were funded. First, the Human Resource Development Council increased the National Training and Reskilling Intake by some 9000 unemployed in construction, manufacturing, logistics, ICT-BPO, agro-industry, renewable energy and the circular economy. These individuals received a monthly benefit of MUR 10,200 for the duration of a 6-month training programme.^{Footnote 15} Second, the Employment Support Scheme for Small and Medium Enterprises was to support 11,000 employees through a monthly benefit of MUR 10,200. Third, Landscope (Mauritius) Ltd was to hire about 2000 unemployed to work with the National Clean-Up Campaign. Fourth, the Air Freight Scheme – incorporated into the Economic Recovery Plan — had two components: namely supervision of the national airline which was under voluntary administration, and support for the export sector. Finally, to support the most vulnerable following a new lockdown, electricity was made free for March and April for individuals on the Social Register of Mauritius or under the National Empowerment Foundation, as well as for low-consuming SMEs, and at a 46% discount for the following 4 months. The Government also established the COVID-19 Solidarity Fund aimed at funding COVID-19 related projects such as financial support to Mauritian residents, and the financing of projects related to the COVID-19 virus and other related health issues. The fund primarily relies on donations from the public and enterprises, however at roughly MUR 500 million it is rather small.

Measuring the effects of the pandemic on the labour market and on living standards

Despite the fast roll-out of government programmes aimed at shielding labour incomes, the question arises as to whether these programmes were effective in mitigating income losses. To provide an answer to this question, we analysed data on self-reported income losses among Mauritian households between May and December 2020 (Figure 6). According to our estimates, on average more than one in three Mauritian households reported an income loss in comparison to their pre pandemic income (in the May, September, October and December round), or in comparison to their previous monthly income (in the July and June round). These income losses were most pronounced at the outset of the COVID-19 pandemic and gradually reduced towards the end of 2020. Whereas 41.7% of households reported experiencing an income loss during May 2020, 30.1% or almost one in three households saw their incomes fall below pre pandemic levels in December 2020 and living standards have continued to worsen for a non-negligible share of the population. First, although employment bounced back in December from the May decline, the sharp contraction in January 2021 hints at the temporary nature of this labour market recovery. Second, for many workers, even individuals who have gone back to work after an interruption, labour income is below pre-pandemic levels. Both individuals who have lost their jobs and displaced workers will need continuing support. In this way the income support schemes, such as SEAS and GWAS, have been useful in maintaining job attachment and protecting some workers from the scarring effects associated with prolonged periods of unemployment, particularly worrisome among youth.

Figure 6. Changes in household income, relative to pre-lockdown (May, September, October and December round), to previous month (June and July round).

Source: Based on data from the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

In the May round, the question regarding the change in household income has a reference period the time since the start of the lockdown; in the June and July round, the reference period is the previous month; in the September–December rounds, the reference period is the first months of 2020 before CVOID-19.

To further disentangle these findings, we performed a regression analysis to study the correlates of the probability of reporting an income loss at the household level between May and December 2020. The results are reported in Figure 7.

Figure 7. Correlates of household income loss by time period.Source: Based on data from the Continuous Multi-Purpose Household Survey and the Rapid Continuous Multi-Purpose Household Survey, Statistics Mauritius and World Bank.

The point estimates are based on OLS regression including month and district fixed effects. The dependent variable in all model specifications is a respondents answer to the question whether they have lost their employment due to business closures between September 2020 and the first quarter of 2021.

In line with the informal sector’s observed decline in activity and limited coverage from government programmes (see Supplemental Table A5), it is those households with a larger share of household members working informally that were more likely to experience income loss. Similarly, those households in which one or more members lost their employment during the reference period were at greater risk of experiencing lost income. Households with more members engaged in financial, health services and science/technology sectors were less likely to report a negative income shock, possibly due to the relatively easy switch to a remote working environment. Relying on previous surveys, our results show that workers in the services sector, especially in modern services (such as information and communication, professional, financial, insurance and real estate activities, scientific and technical activities, administrative and support service activities) are more likely to access computers and therefore potentially perform their work from home (Figure 8a). The skill level of jobs also has an affect with the share of workers able to use computers employed in high (96.3%) and medium (59.4%) skill jobs considerably larger than the 30.8% of workers in low skill jobs (Figure 8b).

Figure 8. Share of the employed able to use computer by sector and skill level first quarter 2020: (a) by sector and (b) by skill level.

Source: Based on data from the 2020 Continuous Multi-Purpose Household Survey, Statistics Mauritius.

Interestingly, while female-headed households represented 9.4% of all households in our sample, they were less likely to be affected by income loss. This effect cannot be attributed to better coping strategies but is likely due to such households being largely dependent on government transfers rather than income from labour, prior to the outbreak of the pandemic (World Bank, 2017).

Although we do not have direct information on the distributional consequences of these income losses, our results indicate that the crisis has impacted those sectors, occupations and informal jobs in which workers from poor and low-income households are concentrated. For example, according to prior estimates, the manufacturing, trade and construction sectors combined employed 39% of all workers but 50% of the poor. Additionally, 66% of the poor worked informally, disproportionately more than in the informal share of the workforce overall (31%). At the same time, it is households exposed to these sectors that are experiencing the largest reported income losses.

The limited availability of home-based work among workers in traditional sectors such as agriculture and manufacturing and in low skill jobs, and the limited access to the internet among the most disadvantaged households, have likely aggravated these adverse economic impact on vulnerable households. Synthesising our results, our findings support the notion that the pandemic has undermined some of the progress achieved in poverty reduction and shared prosperity in recent years.