Did a successful fight against COVID-19 come at a cost? Impacts of the pandemic on employment outcomes in Vietnam

Hai-Anh H Dang; Cuong Viet Nguyen; Calogero Carletto

doi:10.1016/j.worlddev.2022.106129

. 2022 Oct 21;161:106129. doi: 10.1016/j.worlddev.2022.106129

Did a successful fight against COVID-19 come at a cost? Impacts of the pandemic on employment outcomes in Vietnam

Hai-Anh H Dang ^a,^b,^c,^f,^g,^⁎, Cuong Viet Nguyen ^b,^d,^e, Calogero Carletto ^a

PMCID: PMC9584868 PMID: 36284738

Abstract

Despite its low middle-income status, Vietnam has been widely praised for its success in the fight against early waves of the COVID-19 pandemic, with a low mortality rate of approximately 100 deaths out of a population of less than 100 million by the end of 2020. We add to the emerging literature on COVID-19 effects on the labor market for poorer countries by analyzing rich individual-level data from Vietnam’s Labor Force Surveys spanning 2015 to 2020. We find post-pandemic increases in unemployment and temporary layoff rates alongside decreases in employment quality. Monthly wages declined even as the proportion of workers receiving below-minimum wages substantially increased, contributing to sharply rising wage inequality. Our findings suggest that more resources should be allocated to protect vulnerable workers, especially as the pandemic continues to cause increasingly severe damage to the global economy.

Keywords: COVID-19, Employment, Wage inequality, Differences-in-differences, RDD, Vietnam

1. Introduction

The ongoing COVID-19 pandemic has wreaked havoc on rich and poor economies alike around the world. Yet, despite its modest status as a low middle-income country, Vietnam has received strong praise for its early fight against the pandemic, which outperformed richer countries with far more developed medical systems (Huynh, 2020, Trevisan et al., 2020, Hartley et al., 2021). In particular, the country’s strict lockdown measures such as banning all commercial flights into and out of the country, rigorous quarantines, social distancing, and stay-at-home orders were regarded as effective and were strongly supported by the public. As a result, while many countries were still grappling with the outbreak, Vietnam mostly had the pandemic in check in 2020. Tracking data from John Hopkins University suggest that by the end of the year, the country registered an extremely low fatality rate of only 78 deaths, an impressive feat given its population size of slightly more than 96 million.1

But did this preliminary success come at a cost to Vietnam’s labor force? How did the COVID-19 pandemic affect the country’s employment outcomes? Which population subgroups and sectors were most impacted? In particular, were low-income workers affected at higher rates relative to higher-income workers? We seek answers to these questions in order to provide relevant evidence to policy makers in Vietnam who are eager to capitalize on the country’s preliminary medical successes to speed up economic growth, but face the daunting challenge of fighting subsequent pandemic waves as COVID-19 continues to ricochet around the world. The lessons learned from Vietnam’s experience are also relevant for other lower-income countries that plan to implement similarly strict lockdown measures.2

A cursory look at the Vietnam Labor Force Surveys (LFSs) between 2015 and 2020 offers some indication as to the adverse effects of the pandemic. Plotting the average employment outcomes by quarters, Fig. 1 suggests that the average number of working hours during the last seven days remains similar to those in previous years. Yet, both the unemployment rate and the proportion of workers working for less than the minimum wages in Quarters 2 to 4 in 2020 are higher than figures for the corresponding quarters in previous years.3 In fact, the temporary layoff rate sharply increased thirtyfold to 3.1 % in Quarter 2 of 2020. The average monthly wages in the second and the fourth quarters of 2020 were also lower than the corresponding figures for 2019.

As a preview of our findings, employing more rigorous analysis using the difference-in-differences (DD) econometric model, we find that the unemployment and temporary layoff rates increased after the pandemic outbreak in late March 2020. The quality of employment, as measured by wage jobs, jobs with contracts, and formal jobs, was also reduced. Compared to the first quarter of 2020, workers’ monthly wages decreased by 11 % in the second quarter, 7.2 % in the third quarter, and 8.2 % in the fourth quarter. Informal household workers and foreign direct investment (FDI) sector workers were more affected than public sector workers, and workers in the transportation and tourism sectors were most heavily affected. Most worryingly, the proportion of workers working below minimum wages increased by 32 %, fueling significant increases in wage inequality. Some evidence suggests that provinces with greater openness to the global economy witnessed weaker pandemic effects. Further analysis that exploits both timing and geographical variations with lockdown policies, using the regression discontinuity design (RDD) model in combination with difference-in-differences (RDD-DD) and triple differences (RDD-DDD) models, points to lockdown policies as being the primary channel of pandemic impacts.

Beyond offering the first comprehensive analysis of the impacts of COVID-19 on labor outcomes for Vietnam, our study makes several new contributions to the literature. First, we add to the emerging literature on the impacts of the pandemic on labor outcomes in a lower-income country setting. We analyze large-scale, nationally representative, annual LFS data, which span the five years preceding the year of the pandemic and average more than 600,000 households per survey year. The large samples of official labor data before and after the pandemic offer us a rare opportunity to employ rigorous econometric models for analysis.

Indeed, a large number of studies generally find negative pandemic effects on employment in high-income countries (e.g., Adams-Prassl et al., 2020, Coibion et al., 2020, Gupta et al., 2020, Albanesi and Kim, 2021, Dang and Nguyen, 2021), but far fewer studies exist for poorer countries. Since the differences in labor market institutions and available government budgets can vary widely between richer and poorer countries, the policy responses that are effective in high-income countries may not be applicable for poorer countries.

Furthermore, while some early studies for poorer countries usefully document declines in employment after the pandemic, they mostly rely on smaller phone survey samples (e.g., Egger et al., 2021, Khamis et al., 2021, Mahmud and Riley, 2021, Bundervoet et al., 2022). Very few studies analyze large-scale official national labor surveys. Phone survey data are prone to various sampling issues, such as low response rates and under-coverage. They also tend to have shorter questionnaires with far fewer variables than the typical household survey, and so typically do not allow for the rigorous and comprehensive analysis that can be implemented with LFS data.4 To our knowledge, Deshpande (2020) is the only study that analyzes post-pandemic employment using large-scale nationally representative household survey data. Exploring the effects of the pandemic on gender-based differences in employment between April and August 2020, Deshpande (2020) found that women reported higher unemployment levels than men after the first wave of the outbreak, and incomes in the rural sector declined for both genders.

Second, by leveraging large-scale LFS data spanning 2015 to 2020 with new district-level minimum wage data that we manually compile, we are able to more granularly explore the heterogeneous effects of the pandemic on different population subgroups by gender, age, education level, and income level, inter alia. We can also examine disaggregated effects for employment industries and wage quintiles and further map out the effects geographically for different provinces across the country.

Finally, we analyze a wide range of employment indicators, including unemployment, temporary layoffs, labor market participation, employment with labor contract and social insurance, working hours, and monthly wages. We also study the effects of the pandemic on the proportion of below-minimum wage workers and offer new analysis on wage inequality. It has been observed that pandemic-induced labor practices in richer countries (such as home-based work) turn out to be mostly unfavorable to lower-income households and can even exacerbate pre-existing inequalities (Bonacini et al., 2021, Papageorge et al., 2021). Yet, while protecting low-wage workers with minimum wages is among the key labor policies in most countries, to our knowledge, the effects of the pandemic on wage inequality in poorer countries has received little, if any, attention.

This paper consists of six sections. We describe the country background and data in Section 2, followed by a discussion of our estimation methods in Section 3. In Section 4, we test the assumptions underlying our analytical methods (Section 4.1) and provide the estimation results (Section 4.2) and analysis of potential mechanism (Section 4.3), along with various robustness checks (Section 4.4) and heterogeneity analysis (Section 4.5). We offer further analysis on low-wage workers and wage inequality in Section 5 and conclude in Section 6.

2. Country background and data description

2.1. Country background

The first wave of COVID-19 began in Vietnam in late March 2020, with around 10 infection cases detected per day (MoH, 2020). Despite the low case number, Vietnam closed its international borders on March 22, 2020 and imposed a strict nationwide lockdown in April 2020.5 The lockdown was implemented in all 63 provinces, of which 27 provinces applied a 15-day lockdown and the remaining provinces applied a lockdown of 20 to 30 days. The lockdowns were successful, resulting in no new cases by April 2020 and the subsequent resumption of all economic activities. Yet, after three months of no community transmission, the pandemic's second wave began in Da Nang – a major city in central Vietnam (Djalante et al., 2020). A second lockdown was implemented in Da Nang for 6 weeks, between July 27 and September 4, 2020, and social distancing was imposed within several neighboring provinces during this period.

However, the strong lockdown measures against the pandemic were costly for the economy. Despite achieving an impressive, world-leading GDP growth rate of 2.9 percent in 2020, the country’s economic growth in this year was still less than half of rates as high as 6 percent in preceding years (GSO, 2021a). Official statistics from the Government of Vietnam suggests that about 32.1 million people aged 15 and over nationwide were negatively affected by COVID-19 (GSO, 2021b). Various rapid assessment studies based on online or phone surveys implemented between April and September 2020 – the months immediately following the onset of the pandemic – point to pandemic-induced income losses and further suggest that unemployed and informal workers (i.e., those without labor contracts) were the most significantly affected (UN Women, 2020, Yang et al., 2020, Dang et al., 2021, Do et al., 2021). Consequently, the government has been pursuing the dual targets of both containing the pandemic and maintaining economic growth.

2.2. Data description

We analyze Vietnam Labor Force Surveys (LFSs) conducted between 2015 and 2020. LFS data are the official source of labor statistics in Vietnam, collected annually by the country's General Statistics Office (GSO). The LFSs rely on a two-stage stratified cluster design and contain 126 strata comprising of urban and rural areas in 63 provinces throughout the country.6 The surveys are nationally representative on a quarterly basis and at the urban/ rural and provincial levels. The sample size is equally allocated throughout the year, with around one-twelfth of the sampled households being surveyed each month. Aside from collecting basic individual demographic information, the LFSs collect detailed data on employment and wages for people aged 15 and older as well as data on unemployment. Our estimation sample sizes range between 600,000 and more than 620,000 observations for each year in the period 2015–2020.

Minimum wages are adjusted annually and represent an important labor policy issue in Vietnam.7 As such, for the period 2015–2020, we manually collect minimum wage data from the Government of Vietnam’s annual Decrees on minimum wages for each of the approximately 700 districts, and merge these data with the LFS data. These combined data allow us to compute the proportion of workers receiving wages below the minimum wages in their residence district. The nominal minimum wages have been raised annually, with year-on-year increases of 15 % in 2014, 12 % in 2016 and 6 % in 2020 (see Fig. A.3, Appendix A). The nominal minimum wages were not increased in 2021, implying that the real minimum wages decreased slightly in 2021 because of inflation. However, as we analyze the LFS data for the period 2015–2020, the 2021 decrease in the real minimum wage does not pose a concern in this study.

We examine the impacts of the COVID-19 pandemic on a wide range of employment outcomes including unemployment, temporary layoffs, the number of working hours during the last seven days, monthly wages, wages below the minimum wages, and whether workers have a wage job, a job with a contract, or a formal job (that is, a job with social insurance). Table A.1 presents the summary statistics of the outcome variables for the 2015–2020 period. The country’s unemployment rate is low and hovers around 2 % in recent years, possibly because of a large number of people working in the informal sector and the agricultural sector (Demombynes and Testaverde, 2018).8 The proportion of workers with a wage job increased from 39.5 % in 2015 to 48.4 % in 2020. In 2020, wage workers with contracts accounted for 30 % of the workforce, while wage workers with a formal job accounted for 26.7 %.9

3. Estimation method

Since the COVID-19 pandemic occurred in late March 2020 and Vietnam subsequently imposed its first national lockdown in April 2020, any pandemic-induced negative effects on employment and incomes would have occurred starting from Quarter 2 in 2020.10 We estimate the effects of the pandemic on employment outcomes in Vietnam, using a difference-in-differences (DD) econometric model that compares the differences in outcomes between Quarter 1 and the other quarters in 2020 with those averaged over the preceding five years.

Specifically, the observed difference in individuals’ employment outcomes between Quarter 1 and Quarter 2 of 2020 can be expressed as the sum total of the pandemic effects and the seasonal (time) effects as follows11 .

Δ Y = E (Y_{Q 2}^{2020}) - E (Y_{Q 1}^{2020}) = Δ Y_{Covid}^{2020} + Δ Y_{Time}^{2020}

(1)

where $E (Y_{Q 1}^{2020})$ and $E (Y_{Q 2}^{2020})$ are respectively the expected outcomes of individuals in the first and second quarters in 2020, and $Δ Y_{Covid}^{2020}$ and $Δ Y_{Time}^{2020}$ are respectively the COVID-19 effects and seasonal effects. We cannot observe these effects separately. However, assuming that the seasonal effects in 2020 are similar to those in previous years, we can use the latter to substitute for the former. More specifically, we assume.

Δ Y_{Time}^{2020} = E (Y_{Q 2}^{2015 - 2019}) - E (Y_{Q 1}^{2015 - 2019})

(2)

where $E (Y_{Q 1}^{2015 - 2019})$ and $E (Y_{Q 2}^{2015 - 2019})$ are the expected outcomes of the first and second quarters averaged over the past five years from 2015 to 2019. In these years, there were no economic shocks between the first and second quarters; consequently, the averaged differences in the employment outcomes over this period can capture the seasonal effects. Averaging pre-pandemic outcomes over the five preceding years also helps remove fluctuations and provide better comparison, but for robustness checks, we present estimates using any single year in these five years.

Substituting (2) into (1), we obtain.

E (Y_{Q 2}^{2020}) - E (Y_{Q 1}^{2020}) = Δ Y_{Covid}^{2020} + E (Y_{Q 2}^{2015 - 2019}) - E (Y_{Q 1}^{2015 - 2019})

(3)

and after rearranging the terms, we obtain

Δ Y_{Covid}^{2020} = [E (Y_{Q 2}^{2020}) - E (Y_{Q 1}^{2020})] - [E (Y_{Q 2}^{2015 - 2019}) - E (Y_{Q 1}^{2015 - 2019})]

(4)

Equation (4) is a DD estimator, in which the first differences are between the second quarters and the first quarters (i.e., comparing the terms within the two square brackets), and the second differences are between 2020 and the preceding five years, 2015–2019 (i.e., comparing the two square brackets).

More generally, we can extend Equation (4) to examine the pandemic impacts in the other quarters.

Δ Y_{Covid}^{2020} = [E (Y_{Qj}^{2020}) - E (Y_{Q 1}^{2020})] - [E (Y_{Qj}^{2015 - 2019}) - E (Y_{Q 1}^{2015 - 2019})]

(5)

where j indicates the quarter of the year, j = 2, 3, or 4.

Pooling all the quarters together, we obtain the estimating regression for Equation (5).

y = β_{0} + β_{1} (PY . Q_{2}) + β_{2} (PY . Q_{3}) + β_{3} (PY . Q_{4}) + β_{4} Q_{2} + β_{5} Q_{3} + β_{6} Q_{4} + β_{7} P Y + B^{'} X + σ_{d} + τ_{t} + ε_{dt}

(6)

where y is an employment outcome for individuals, PY is a dummy variable indicating 2020, the year of the pandemic, and Q₂, Q₃, and Q₄ are the dummy variables corresponding to Quarter 2, Quarter 3, and Quarter 4, with Quarter 1 being the reference quarter. The pandemic effects on employment outcomes in these quarters are measured by $β_{1}$ , $β_{2}$ , and $β_{3}$ – the coefficients of the interaction terms between PY and Q₂, Q₃ _, and Q₄. X is the matrix of control variables including age, gender, and education, and ε is the error term. We also include in Equation (6) the district fixed effects ( $σ_{d}$ ) and the year fixed effects ( $τ_{t}$ ) to control for unobserved factors that occur in the same district or the same year. The summary statistics of the control variables are presented for each year in Table A.2 in Appendix A.

As LFSs are collected on a monthly basis, we can further estimate the immediate pandemic impacts on employment outcomes by month. These monthly impacts start from April 2020 and are estimated in comparison to Quarter 1, 2020. Specifically, we can replace the dummy variables for Quarters 2 to 4 in Equation (5) with the dummy variables indicating the months.

Δ Y_{Covid}^{2020} = [E (Y_{Mk}^{2020}) - E (Y_{Q 1}^{2020})] - [E (Y_{Mk}^{2017 - 2019}) - E (Y_{Q 1}^{2017 - 2019})]

(7)

where k indicates the month of the year, k = 4, 5,.., 12. Similarly to Equation (6), we can estimate Equation (7) with the following DD regression.

y = α + \sum_{k = 4}^{12} θ_{k} P Y . M_{k} + \sum_{k = 4}^{12} γ_{k} M_{k} + φ P Y + Λ^{'} X + μ_{d} + ρ_{t} + v_{dt}

(8)

where $θ_{k}$ are the coefficients of interest.

While Equation (8) can broadly capture the net changes with labor outcomes before and after the pandemic (that is, the total economy-wide effects that are pandemic-induced), it does not pinpoint any specific mechanism that causes these changes. One such key mechanism could be the lockdown policy implemented by the Government of Vietnam.12 Consequently, to estimate the local effects caused by the pandemic-induced lockdowns, we offer a multi-pronged approach that examines more variations regarding the lockdowns along the time dimension, the geographical dimension, and their combinations.

The first alternative is to use the sharp regression discontinuity design (RDD) model. We use the lockdown date of April 2020 as the cutoff, and we use a bandwidth of 9 months on either side of this cutoff (that is, July 2019 to March 2020 on the left, and April 2020 to December 2020 on the right). Put differently, we take month as the conditioning (assignment) variable. We estimate the following equation

y = δ_{0} + δ_{1} L M + δ_{2} M o n t h + δ_{3} M o n t h . L M + Δ^{'} X + ϛ_{d} + ϕ_{t} + ξ_{dt}

(9)

where $Month$ equals 0 for April 2020 and ranges from −9 (July 2019) to 8 (December 2020). The treatment variable is the national lockdown, which equals 1 for the months starting from April 2020 and 0 otherwise (i.e., $LM$ equals 1 if $Month \geq 0$ , and 0 otherwise). The local effects of the April 2020 lockdown are estimated by $δ_{1}$ .

However, Equation (9) may capture not only the lockdown effects but also the seasonal (or any unobserved macro-economic) effects in 2020. To examine these seasonal effects, we can estimate the same model in Equation (9) using data from prior to 2020 (i.e., the 2015–2019 period), which can serve as the placebo test for the RDD model. If seasonal effects exist, we can combine a DD estimation strategy with the RDD model in Equation (9) for more robust analysis.13 Specifically, we estimate the following RDD-DD regression.

y = λ_{0} + λ_{1} L M . C O V I D + λ_{2} M o n t h . C O V I D + λ_{3} M o n t h + λ_{4} M o n t h . L M . C O V I D + λ_{5} L M + λ_{6} M o n t h . L M + λ_{7} C O V I D + Γ^{'} X + Θ^{'} X . C O V I D + o_{d} + ϱ_{t} + ∊_{dt}

(10)

where COVID is a dummy variable that equals 1 for the period July 2019 to December 2020, and 0 for the months before July 2019. The coefficient of interest is $λ_{1}$ .

For the second alternative, we make use of the fact that a major central city in Vietnam – Da Nang – experienced a second lockdown from July to September 2020. To examine whether this second lockdown affected Da Nang beyond the impacts of the first lockdown from the DD model, we include the interaction between a dummy variable for Da Nang and the treatment variables. Building on Equation (6), we estimate the following triple differences (DDD) model, in which we combine Quarters 2 to 4 into a dummy Q₂₄ to measure the combined effect of the pandemic in three quarters.

y = α_{0} + α_{1} (PY . Q_{24}) + α_{2} (PY . Q_{24} . D a N a n g) + α_{3} Q_{24} + α_{4} P Y + C^{'} X + s_{d} + r_{t} + e_{dt}

(11)

where $Q_{24}$ is a dummy indicating Quarter 2 to 4, and $DaNang$ is the dummy variable indicating the city of Da Nang. Note that the variable DaNang does not appear on its own in Equation (11), as we already control for the district fixed effects in this equation. The coefficient $α_{2}$ on the triple interaction term $PY . Q_{24} . D a N a n g$ measures the difference in the lockdown effects between Da Nang and other provinces and cities in Vietnam. Put differently, $α_{2}$ measures the impacts of the second lockdown on Da Nang.

Finally, we can combine these alternatives to better exploit both the timing and geographical variations related to the lockdowns for identification. We further interact DaNang and the treatment variables in Equations (9), (10) to estimate the following RDD-DD equation.

y = π_{0} + π_{1} L M + π_{2} L M . D a N a n g + π_{3} M o n t h + π_{4} M o n t h . L M + Δ^{'} X + ϛ_{d} + ϕ_{t} + ξ_{dt}

(12)

and the RDD-DDD equation

y = ρ_{0} + ρ_{1} L M . C O V I D + ρ_{2} L M . C O V I D . D a N a n g + ρ_{3} M o n t h . C O V I D + ρ_{4} M o n t h + ρ_{5} M o n t h . L M . C O V I D + ρ_{6} L M + ρ_{7} M o n t h . L M + ρ_{8} C O V I D + Γ^{'} X + Θ^{'} X . C O V I D + o_{d} + ℓ_{t} + m_{dt}

(13)

$π_{2}$ and $ρ_{2}$ are respectively the coefficients of interest in Equations (11), (12). While Equations (10), (11), (12) estimate the effects of the second lockdown on Da Nang alone (as only this city underwent a second lockdown in Vietnam during 2020), the results can provide supportive evidence for the lockdowns as a key channel for the adverse impacts of the pandemic on employment and wage inequality.

4. Empirical results

4.1. Testing assumptions

Our DD estimation strategy relies on two key assumptions. The first assumption is that employment outcomes in Vietnam were not affected by the pandemic in Quarter 1 of 2020 (such that this quarter can represent the reference quarter in Equation (5)). This is a reasonable assumption, as the pandemic occurred in late March 2020 (and lockdown measures occurred after that). Indeed, plotting the employment outcomes on a quarterly basis for the period 2015–2020, Fig. 1 shows that, compared to Quarter 1 in the preceding years, Quarter 1 of 2020 generally has a lower unemployment rate, higher proportions of wage jobs and of workers with labor contracts and social insurance, a higher number of working hours, a higher average wage, and a lower proportion of below-minimum wage workers. The only exception is the temporary layoff rate, but this rate is very low, at <0.1 % for all the first quarters studied.

To formally test this assumption, we compare changes in the employment outcomes between Quarter 4 of 2019 and Quarter 1 of 2020 with similar changes between the corresponding quarters of the preceding years. We restrict the sample to the first and fourth quarters and construct a dummy variable (denoted by Year 2019–2020 in Table A.3) which equals 1 for the fourth quarter of 2019 and the first quarter of 2020, and 0 otherwise. We regress the employment outcomes on this variable, a dummy variable for the first quarters, the interaction term between these two variables, and other control variables. The interaction term represents the pandemic effects in the first quarter of 2020 and follows the same DD strategy as in Equation (6). The regression results, reported in Table A.3, indicate that the lockdown has statistically insignificant effects in Quarter 1 of 2020.

The second assumption is the standard “parallel trend” assumption for the DD model, which requires that in the absence of the pandemic, the changes in outcomes between Quarter 1 and the other quarters of 2020 are similar to the corresponding changes between Quarter 1 and the other quarters of the preceding years. To visually examine this assumption, we plot in Fig. A.2 the employment outcomes of Quarter 1 against those for the remaining three quarters over the six years studied. If the parallel trend assumption is satisfied, we should see parallel lines that represent the outcomes in the pre-pandemic years. Indeed, Fig. A.2 shows that the lines are roughly parallel during 2015–2019, but either cross-cut each other (Panel A, B, C, and F respectively for unemployment, temporary layoffs, and the proportions of workers having a wage job or working below the minimum wages) or diverge (Panel E, monthly wages) in 2020. For workers with a labor contract or formal job, the lines closely track each other during 2015–2019 but intersect in 2020 (Panels G and H). These results support the parallel trend assumption.

4.2. Estimated pandemic impacts

Table 1 reports the DD regressions of employment outcomes using the LFS data from 2015 to 2020 (using Equation (6)). Aside from working hours, the interaction terms between the pandemic year (PY) and Quarters 2 to 4 are statistically significant in all the regressions, which suggests that the pandemic negatively impacted employment outcomes for Quarters 2 to 4 of 2020. The estimated impact magnitudes are largely similar.

Table 1.

DD regressions of employment variables.

Explanatory variables	Dependent variables
	Unemployed (yes = 1, no = 0)	Temporary layoff (yes = 1, no = 0)	Having a wage job (yes = 1, no = 0)	Having job with contract (yes = 1, no = 0)	Having a formal job (yes = 1, no = 0)	Log of number of working hours in the last 7 days	Log of monthly wage (wage workers)	Having wage below minimum wage (yes = 1, no = 0)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Quarter 2 * COVID year	0.0098***	0.0305***	−0.0171***	−0.0099***	−0.0076***	−0.0073	−0.1096***	0.0553***
	(0.0012)	(0.0004)	(0.0040)	(0.0015)	(0.0015)	(0.0213)	(0.0027)	(0.0022)
Quarter 3 * COVID year	0.0084***	0.0028***	−0.0128***	−0.0142***	−0.0124***	0.0263	−0.0721***	0.0278***
	(0.0011)	(0.0004)	(0.0041)	(0.0016)	(0.0015)	(0.0212)	(0.0024)	(0.0015)
Quarter 4 * COVID year	0.0090***	0.0004**	−0.0145***	−0.0113***	−0.0100***	0.0313	−0.0819***	0.1008***
	(0.0014)	(0.0002)	(0.0042)	(0.0012)	(0.0011)	(0.0217)	(0.0036)	(0.0027)
Quarter 1	Reference
Quarter 2	−0.0039***	−0.0006***	0.0137***	0.0032***	0.0021***	0.0533**	−0.0216***	−0.0138***
	(0.0012)	(0.0001)	(0.0044)	(0.0007)	(0.0007)	(0.0215)	(0.0038)	(0.0016)
Quarter 3	−0.0051***	−0.0007***	0.0134***	0.0038***	0.0025***	0.0645***	−0.0027	−0.0262***
	(0.0011)	(0.0001)	(0.0045)	(0.0006)	(0.0007)	(0.0213)	(0.0039)	(0.0018)
Quarter4	−0.0075***	−0.0008***	0.0281***	0.0104***	0.0087***	0.0876***	0.0010	−0.0321***
	(0.0014)	(0.0001)	(0.0046)	(0.0008)	(0.0007)	(0.0219)	(0.0049)	(0.0019)
COVID year	−0.0004	−0.0004**	0.0793***	0.0293***	0.0347***	−0.0267	0.2746***	−0.0606***
	(0.0010)	(0.0002)	(0.0081)	(0.0037)	(0.0038)	(0.0191)	(0.0056)	(0.0028)
Male (male = 1, female = 0)	0.0003	0.0003*	0.0948***	−0.0298***	−0.0383***	0.0764***	0.1770***	−0.0507***
	(0.0008)	(0.0001)	(0.0038)	(0.0037)	(0.0036)	(0.0035)	(0.0050)	(0.0039)
Age	−0.0056***	0.0000**	−0.0014**	−0.0001	0.0039***	0.0219***	0.0590***	−0.0202***
	(0.0003)	(0.0000)	(0.0006)	(0.0005)	(0.0006)	(0.0010)	(0.0015)	(0.0010)
Age squared	0.0001***	−0.0000***	−0.0001***	−0.0001***	−0.0001***	−0.0003***	−0.0008***	0.0003***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education	Reference
Primary education	0.0002	0.0001	−0.1092***	−0.0998**	−0.0966**	0.0129***	−0.0478	−0.0124
	(0.0016)	(0.0001)	(0.0249)	(0.0371)	(0.0366)	(0.0039)	(0.0429)	(0.0077)
Lower-secondary education	0.0000	0.0001	−0.1220***	−0.0709	−0.0730*	0.0191***	0.0016	−0.0271***
	(0.0018)	(0.0001)	(0.0280)	(0.0416)	(0.0412)	(0.0045)	(0.0462)	(0.0084)
Upper-secondary education	0.0044*	−0.0001	−0.0396	0.0965**	0.0789*	0.0453***	0.0682	−0.0375***
	(0.0022)	(0.0003)	(0.0305)	(0.0455)	(0.0448)	(0.0050)	(0.0489)	(0.0092)
Post-secondary education	0.0177**	−0.0006***	0.2100***	0.4620***	0.4612***	−0.0097	0.3262***	−0.0666***
	(0.0066)	(0.0002)	(0.0485)	(0.0613)	(0.0593)	(0.0087)	(0.0481)	(0.0113)
Urban (urban = 1, rural = 0)	0.0084***	0.0006**	0.0526***	0.0477***	0.0427***	0.0660***	0.0520***	0.0003
	(0.0010)	(0.0002)	(0.0089)	(0.0107)	(0.0104)	(0.0049)	(0.0082)	(0.0024)
Year fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.1439***	0.0008***	0.5935***	0.3305***	0.2066***	3.2270***	7.2432***	0.4992***
	(0.0060)	(0.0002)	(0.0318)	(0.0430)	(0.0417)	(0.0277)	(0.0532)	(0.0251)
Observations	2,759,355	2,759,355	2,759,355	2,759,355	2,759,355	2,682,379	1,124,748	1,124,748
R-squared	0.031	0.011	0.248	0.334	0.328	0.210	0.346	0.089

Data sample and model	Dependent variables
	Unemployed (yes = 1, no = 0)	Temporary layoff (yes = 1, no = 0)	Having wage job (yes = 1, no = 0)	Having job with contract (yes = 1, no = 0)	Having social insurance (yes = 1, no = 0)	Log of number of working hours in the last 7 days	Log of monthly wage (wage workers)	Having wage below minimum wages (yes = 1, no = 0)
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
RDD estimates using the data sample of LFSs 2019 and 2020	0.0137***	0.0338***	−0.0258***	−0.0193***	−0.0172***	−0.0165	−0.1076***	0.0180
RDD estimates using the data sample of LFSs 2019 and 2020	(0.0025)	(0.0095)	(0.0061)	(0.0045)	(0.0042)	(0.0296)	(0.0233)	(0.0147)
RDD-DD estimates using the data sample of LFSs 2017 to 2020	0.0162***	0.0347***	−0.0284***	−0.0207***	−0.0173***	−0.0917***	−0.0745***	0.0249*
RDD-DD estimates using the data sample of LFSs 2017 to 2020	(0.0024)	(0.0096)	(0.0058)	(0.0048)	(0.0040)	(0.0240)	(0.0233)	(0.0147)
RDD-DD estimates using the data sample of LFSs 2015 to 2020	0.0169***	0.0346***	−0.0264***	−0.0188***	−0.0159***	−0.0783***	−0.0701***	0.0245*
RDD-DD estimates using the data sample of LFSs 2015 to 2020	(0.0022)	(0.0096)	(0.0058)	(0.0056)	(0.0054)	(0.0188)	(0.0201)	(0.0145)

FGT and inequality indexes	Index with lockdowns (observed)	Index without lockdowns (estimated)	Difference	Difference (%)
	(1)	(2)	(3)=(1)-(2)	(4)=(3)/(2)
FGT indexes
P0	10.2756***	7.7637***	2.5119***	32.36***
	(0.0803)	(0.0718)	(0.0444)	(0.67)
P1	0.0315***	0.0249***	0.0065***	26.23***
	(0.0003)	(0.0003)	(0.0002)	(1.06)
P2	0.0151***	0.0119***	0.0032***	27.12***
	(0.0002)	(0.0002)	(0.0001)	(1.23)
Inequality indexes
Gini index	0.2514***	0.2402***	0.0112***	4.65***
	(0.0008)	(0.0011)	(0.0008)	(0.33)
Theil L	0.1179***	0.1071***	0.0109***	10.16***
	(0.0007)	(0.0009)	(0.0006)	(0.63)
Theil T	0.1153***	0.1070***	0.0083***	7.78***
	(0.0010)	(0.0012)	(0.0006)	(0.60)
p(90)/p(10)	2.8947***	2.6874***	0.2073***	7.73***
	(0.0084)	(0.0338)	(0.0319)	(1.29)
p(95)/p(5)	5.0934***	4.4866***	0.6068***	13.53***
	(0.0465)	(0.0341)	(0.0367)	(0.84)

Outcomes	2015	2016	2017	2018	2019	2020
	(1)	(2)	(3)	(4)	(5)	(6)
Unemployment rate (%)	1.687	1.328	2.044	1.995	2.047	2.284
	(0.061)	(0.052)	(0.070)	(0.064)	(0.084)	(0.087)
Temporary layoff rate (%)	0.083	0.060	0.077	0.081	0.051	0.861
	(0.015)	(0.008)	(0.007)	(0.009)	(0.005)	(0.043)
Proportion of workers having a wage job (%)	39.5	41.1	41.9	43.1	47.9	48.4
Proportion of workers having a wage job (%)	(0.9)	(0.9)	(0.8)	(0.7)	(0.9)	(0.9)
Proportion of workers having a job with contract (%)	24.8	25.3	25.1	25.6	29.9	30.1
Proportion of workers having a job with contract (%)	(1.0)	(1.0)	(0.8)	(0.8)	(1.1)	(1.1)
Proportion of workers having a formal job (%)	20.8	21.5	21.6	22.6	26.3	26.7
Proportion of workers having a formal job (%)	(0.9)	(0.9)	(0.8)	(0.8)	(1.0)	(1.0)
Number of working hours in the last 7 days	41.2	41.5	40.3	41.6	41.0	41.2
Number of working hours in the last 7 days	(0.3)	(0.3)	(0.2)	(0.2)	(0.3)	(0.2)
Monthly wage of wage workers (thousand VND/month)	5666.2	5885.4	6077.7	6312.7	6928.3	6907.6
Monthly wage of wage workers (thousand VND/month)	(82.3)	(93.8)	(85.9)	(84.0)	(106.2)	(95.0)
% workers with wage below the minimum wages	8.6	9.4	7.7	6.6	4.6	7.3
% workers with wage below the minimum wages	(0.3)	(0.4)	(0.3)	(0.3)	(0.2)	(0.3)

Variables	2015		2016		2017		2018		2019		2020
	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.	Mean	Std. Dev.
Male (male = 1, female = 0)	0.518	0.500	0.516	0.500	0.519	0.500	0.522	0.500	0.532	0.499	0.530	0.499
Age	40.11	13.52	40.55	13.51	40.31	13.54	40.62	13.43	39.82	13.12	40.28	12.93
Less than primary education	0.140	0.347	0.132	0.339	0.139	0.426	0.135	0.342	0.148	0.355	0.113	0.317
Primary education	0.233	0.423	0.230	0.421	0.226	0.418	0.221	0.415	0.212	0.409	0.213	0.409
Lower-secondary education	0.313	0.464	0.315	0.465	0.309	0.462	0.311	0.463	0.292	0.455	0.305	0.460
Upper-secondary education	0.198	0.398	0.201	0.401	0.203	0.402	0.204	0.403	0.201	0.401	0.216	0.412
Post-secondary education	0.116	0.320	0.121	0.326	0.124	0.153	0.128	0.334	0.147	0.354	0.153	0.360
Urban area (urban = 1, rural = 0)	0.318	0.466	0.319	0.466	0.322	0.467	0.326	0.469	0.331	0.471	0.337	0.473
Number of observations	465,570		467,931		471,974		468,156		445,595		440,129

Explanatory variables	Full sample		Informal sector		Formal sector
	Below minimum wages (yes = 1, no = 0)	Below minimum wages (main job) (yes = 1, no = 0)	Below minimum wages (yes = 1, no = 0)	Below minimum wages (main job) (yes = 1, no = 0)	Below minimum wages (yes = 1, no = 0)	Below minimum wages (main job) (yes = 1, no = 0)
	(1)	(2)	(3)	(4)	(5)	(6)
Quarter 2 * COVID year	0.0553***	0.0660***	0.0832***	0.1065***	0.0309***	0.0354***
	(0.0022)	(0.0022)	(0.0027)	(0.0030)	(0.0026)	(0.0021)
Quarter 3 * COVID year	0.0278***	0.0415***	0.0415***	0.0700***	0.0100***	0.0179***
	(0.0015)	(0.0020)	(0.0018)	(0.0031)	(0.0025)	(0.0020)
Quarter 4 * COVID year	0.1008***	0.1062***	0.1305***	0.1377***	0.0683***	0.0793***
	(0.0027)	(0.0024)	(0.0032)	(0.0039)	(0.0024)	(0.0017)
Quarter 1	Reference
Quarter 2	−0.0138***	−0.0083***	−0.0235***	−0.0159***	−0.0053***	−0.0036**
	(0.0016)	(0.0019)	(0.0018)	(0.0029)	(0.0019)	(0.0014)
Quarter 3	−0.0262***	−0.0213***	−0.0405***	−0.0368***	−0.0112***	−0.0090***
	(0.0018)	(0.0021)	(0.0027)	(0.0034)	(0.0022)	(0.0016)
Quarter4	−0.0321***	−0.0266***	−0.0487***	−0.0459***	−0.0133***	−0.0104***
	(0.0019)	(0.0022)	(0.0024)	(0.0035)	(0.0018)	(0.0014)
COVID year	−0.0606***	−0.0781***	−0.0922***	−0.1336***	−0.0224***	−0.0298***
	(0.0028)	(0.0029)	(0.0038)	(0.0052)	(0.0034)	(0.0026)
Male (male = 1, female = 0)	−0.0507***	−0.0619***	−0.1341***	−0.1691***	−0.0040**	−0.0014
	(0.0039)	(0.0036)	(0.0066)	(0.0063)	(0.0015)	(0.0011)
Age	−0.0202***	−0.0256***	−0.0236***	−0.0274***	−0.0071***	−0.0097***
	(0.0010)	(0.0009)	(0.0011)	(0.0008)	(0.0005)	(0.0007)
Age squared	0.0003***	0.0004***	0.0003***	0.0004***	0.0001***	0.0001***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education	Reference
Primary education	−0.0124	0.0058	−0.0289***	−0.0278***	0.0058**	0.0267***
	(0.0077)	(0.0137)	(0.0045)	(0.0040)	(0.0027)	(0.0051)
Lower-secondary education	−0.0271***	−0.0130	−0.0385***	−0.0358***	0.0031	0.0209***
	(0.0084)	(0.0150)	(0.0048)	(0.0045)	(0.0031)	(0.0052)
Upper-secondary education	−0.0375***	−0.0381**	−0.0367***	−0.0342***	0.0047	0.0155***
	(0.0092)	(0.0160)	(0.0050)	(0.0049)	(0.0031)	(0.0052)
Post-secondary education	−0.0666***	−0.0993***	−0.0553***	−0.0865***	−0.0104**	−0.0129**
	(0.0113)	(0.0178)	(0.0082)	(0.0091)	(0.0042)	(0.0058)
Urban (urban = 1, rural = 0)	0.0003	−0.0148***	0.0011	−0.0043	0.0013	−0.0082***
	(0.0024)	(0.0038)	(0.0035)	(0.0040)	(0.0016)	(0.0015)
Year fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.4992***	0.6286***	0.6876***	0.8224***	0.1584***	0.2082***
	(0.0251)	(0.0246)	(0.0279)	(0.0228)	(0.0130)	(0.0147)
Observations	1,110,764	1,110,764	531,635	531,635	579,116	579,116
R-squared	0.089	0.113	0.128	0.145	0.033	0.037
Note: This table presents estimates of the effects of COVID-19 on employment using the DD method. The effects are estimated by the interaction between the second, third quarter and fourth and the 2020 year.Robust standard errors in parentheses (corrected for sampling weight and clustered at the district level and year-by-quarter level) . * p < 0.01, p < 0.05, * p < 0.1.

Explanatory variables	Dependent variables
	Having job with long-term contract (yes = 1, no = 0)	Working for firms and organizations (yes = 1, no = 0)	Having a secondary job (yes = 1, no = 0)	Log of number of working hours of the main job (last 7 days)	Log of monthly wages of the main job	Log of hourly wages of the main job	Log of hourly wages of all jobs
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Quarter 2 * COVID year	−0.0099***	−0.0077***	0.0050**	0.0031	−0.1098***	−0.1237***	−0.1192***
	(0.0015)	(0.0015)	(0.0020)	(0.0205)	(0.0031)	(0.0218)	(0.0223)
Quarter 3 * COVID year	−0.0156***	−0.0117***	0.0025	0.0392*	−0.0714***	−0.1175***	−0.1091***
	(0.0017)	(0.0015)	(0.0022)	(0.0205)	(0.0025)	(0.0216)	(0.0222)
Quarter 4 * COVID year	−0.0109***	−0.0099***	−0.0258***	0.0400*	−0.0824***	−0.1287***	−0.1247***
	(0.0012)	(0.0012)	(0.0026)	(0.0211)	(0.0036)	(0.0221)	(0.0227)
Quarter 1	Reference
Quarter 2	0.0019***	0.0034***	−0.0031	0.0518**	−0.0220***	−0.0539**	−0.0549**
	(0.0006)	(0.0008)	(0.0024)	(0.0206)	(0.0041)	(0.0217)	(0.0223)
Quarter 3	0.0015**	0.0032***	−0.0041	0.0610***	−0.0017	−0.0613***	−0.0635***
	(0.0006)	(0.0006)	(0.0025)	(0.0206)	(0.0042)	(0.0218)	(0.0224)
Quarter4	0.0079***	0.0099***	−0.0070**	0.0804***	0.0030	−0.0882***	−0.0904***
	(0.0007)	(0.0007)	(0.0027)	(0.0212)	(0.0051)	(0.0224)	(0.0231)
COVID year	0.0258***	0.0304***	−0.0232***	−0.0510**	0.2833***	0.3222***	0.3212***
	(0.0038)	(0.0035)	(0.0035)	(0.0183)	(0.0058)	(0.0213)	(0.0215)
Male (male = 1, female = 0)	−0.0284***	−0.0241***	0.0055	0.0655***	0.1776***	0.1472***	0.1467***
	(0.0036)	(0.0038)	(0.0035)	(0.0031)	(0.0050)	(0.0048)	(0.0048)
Age	0.0034***	−0.0005	0.0135***	0.0274***	0.0572***	0.0512***	0.0520***
	(0.0005)	(0.0006)	(0.0006)	(0.0011)	(0.0014)	(0.0013)	(0.0012)
Age squared	−0.0001***	−0.0001***	−0.0001***	−0.0004***	−0.0007***	−0.0006***	−0.0006***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education	Reference
Primary education	−0.0966**	−0.0985**	0.0165***	0.0216***	−0.0499	−0.1024*	−0.1002*
	(0.0364)	(0.0365)	(0.0026)	(0.0044)	(0.0430)	(0.0503)	(0.0500)
Lower-secondary education	−0.0720*	−0.0714*	0.0171***	0.0288***	−0.0025	−0.0647	−0.0635
	(0.0408)	(0.0407)	(0.0030)	(0.0048)	(0.0462)	(0.0547)	(0.0542)
Upper-secondary education	0.0813*	0.0917*	−0.0140***	0.0385***	0.0657	0.0325	0.0304
	(0.0446)	(0.0445)	(0.0031)	(0.0054)	(0.0489)	(0.0581)	(0.0577)
Post-secondary education	0.4419***	0.4552***	−0.0251***	−0.0275***	0.3259***	0.3830***	0.3789***
	(0.0609)	(0.0599)	(0.0044)	(0.0092)	(0.0484)	(0.0575)	(0.0575)
Urban (urban = 1, rural = 0)	0.0465***	0.0510***	−0.0538***	0.0295***	0.0602***	0.0701***	0.0644***
	(0.0103)	(0.0108)	(0.0034)	(0.0043)	(0.0089)	(0.0086)	(0.0088)
Year fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.2175***	0.3479***	−0.1884***	3.1911***	7.2652***	2.1339***	2.1241***
	(0.0424)	(0.0419)	(0.0141)	(0.0301)	(0.0509)	(0.0695)	(0.0679)
Observations	2,759,263	2,759,263	2,759,263	2,685,240	1,124,495	1,110,806	1,109,228
R-squared	0.318	0.326	0.153	0.165	0.352	0.341	0.339
Note: This table presents estimates of the effects of COVID-19 on employment using the DD method. The effects are estimated by the interaction between the second, third quarter and fourth and the 2020 year.Robust standard errors in parentheses (corrected for sampling weight and clustered at the district level and year-by-quarter level) . * p < 0.01, p < 0.05, * p < 0.1.

Explanatory variables	Unemployed (yes = 1, no = 0)	Temporary layoff (yes = 1, no = 0)	Having wage job (yes = 1, no = 0)	Having job with contract (yes = 1, no = 0)	Having social insurance (yes = 1, no = 0)	Log of number of working hours in the last 7 days	Log of monthly wage (wage workers)	Having wage below minimum wages (yes = 1, no = 0)
Lockdown	0.0137***	0.0338***	−0.0258***	−0.0193***	−0.0172***	−0.0165	−0.1076***	0.0180
	(0.0025)	(0.0095)	(0.0061)	(0.0045)	(0.0042)	(0.0296)	(0.0233)	(0.0147)
Month variable	−0.0004	0.0001	0.0021**	0.0012*	0.0014**	−0.0057	0.0052***	−0.0013**
	(0.0003)	(0.0000)	(0.0009)	(0.0006)	(0.0006)	(0.0038)	(0.0019)	(0.0005)
Month variable * Lockdown	−0.0004	−0.0059***	0.0009	−0.0001	−0.0007	0.0179***	0.0015	0.0080**
	(0.0004)	(0.0017)	(0.0011)	(0.0008)	(0.0008)	(0.0043)	(0.0049)	(0.0037)
Male (male = 1, female = 0)	−0.0039***	−0.0005*	0.0837***	−0.0438***	−0.0510***	0.0807***	0.1840***	−0.0312***
	(0.0011)	(0.0003)	(0.0067)	(0.0037)	(0.0039)	(0.0059)	(0.0079)	(0.0028)
Age	−0.0056***	−0.0000	−0.0001	0.0004	0.0041***	0.0236***	0.0501***	−0.0132***
	(0.0003)	(0.0000)	(0.0009)	(0.0008)	(0.0008)	(0.0013)	(0.0017)	(0.0009)
Age squared	0.0001***	−0.0000	−0.0001***	−0.0001***	−0.0001***	−0.0004***	−0.0006***	0.0002***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education	Reference
Primary education	−0.0031***	0.0005	−0.0528***	−0.0349***	−0.0260***	−0.0063	0.0082	−0.0037
	(0.0008)	(0.0003)	(0.0055)	(0.0080)	(0.0065)	(0.0045)	(0.0132)	(0.0042)
Lower-secondary education	−0.0024**	0.0007*	−0.0586***	−0.0003	0.0059	0.0080*	0.0530***	−0.0164***
	(0.0010)	(0.0004)	(0.0073)	(0.0088)	(0.0077)	(0.0042)	(0.0136)	(0.0040)
Upper-secondary education	0.0007	0.0017**	−0.0023	0.1545***	0.1496***	0.0307***	0.1112***	−0.0201***
	(0.0014)	(0.0008)	(0.0089)	(0.0077)	(0.0063)	(0.0051)	(0.0156)	(0.0043)
Post-secondary education	0.0074***	−0.0002	0.2522***	0.5363***	0.5437***	−0.0081	0.3315***	−0.0377***
	(0.0016)	(0.0005)	(0.0148)	(0.0166)	(0.0144)	(0.0092)	(0.0188)	(0.0051)
Urban area (urban = 1, rural = 0)	0.0087***	0.0017**	0.0232***	0.0203***	0.0157**	0.0605***	0.0278***	0.0043
	(0.0014)	(0.0007)	(0.0088)	(0.0076)	(0.0069)	(0.0063)	(0.0063)	(0.0032)
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.1435***	0.0004	0.6533***	0.3242***	0.2046***	3.2535***	7.6415***	0.2773***
	(0.0078)	(0.0006)	(0.0283)	(0.0253)	(0.0229)	(0.0373)	(0.0316)	(0.0160)
Observations	662,211	662,220	662,220	662,220	662,220	637,913	290,442	290,442
R-squared	0.035	0.030	0.255	0.375	0.373	0.244	0.351	0.080
Robust standard errors in parentheses * p < 0.01, p < 0.05, * p < 0.1

Explanatory variables	Unemployed (yes = 1, no = 0)	Temporary layoff (yes = 1, no = 0)	Having wage job (yes = 1, no = 0)	Having job with contract (yes = 1, no = 0)	Having social insurance (yes = 1, no = 0)	Log of number of working hours in the last 7 days	Log of monthly wage (wage workers)	Having wage below minimum wages (yes = 1, no = 0)
Lockdown * Sample 2019–2020	0.0162***	0.0347***	−0.0284***	−0.0207***	−0.0173***	−0.0917***	−0.0745***	0.0249*
	(0.0024)	(0.0096)	(0.0058)	(0.0048)	(0.0040)	(0.0240)	(0.0233)	(0.0147)
Lockdown	−0.0025*	−0.0007***	0.0017	0.0009	−0.0002	0.0754***	−0.0333**	−0.0060*
	(0.0014)	(0.0003)	(0.0049)	(0.0035)	(0.0031)	(0.0144)	(0.0136)	(0.0034)
Month variable	0.0007***	0.0001***	−0.0008	−0.0027***	−0.0027***	−0.0139***	0.0063***	0.0016***
	(0.0002)	(0.0000)	(0.0007)	(0.0006)	(0.0005)	(0.0019)	(0.0020)	(0.0005)
Month variable * Lockdown	−0.0016***	−0.0001***	0.0032***	0.0037***	0.0038***	0.0185***	−0.0027	−0.0041***
	(0.0003)	(0.0000)	(0.0009)	(0.0006)	(0.0006)	(0.0024)	(0.0021)	(0.0006)
Male (male = 1, female = 0)	−0.0023***	0.0002*	0.0891***	−0.0386***	−0.0454***	0.0777***	0.1804***	−0.0406***
	(0.0008)	(0.0001)	(0.0056)	(0.0034)	(0.0035)	(0.0047)	(0.0064)	(0.0031)
Age	−0.0060***	0.0000	−0.0001	0.0003	0.0041***	0.0230***	0.0566***	−0.0171***
	(0.0003)	(0.0000)	(0.0008)	(0.0007)	(0.0007)	(0.0010)	(0.0014)	(0.0008)
Age squared	0.0001***	−0.0000	−0.0001***	−0.0001***	−0.0001***	−0.0003***	−0.0007***	0.0002***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education
Primary education	−0.0006	0.0000	−0.0954***	−0.0867***	−0.0810***	0.0071**	−0.0425***	−0.0100***
	(0.0008)	(0.0001)	(0.0071)	(0.0122)	(0.0116)	(0.0033)	(0.0146)	(0.0035)
Lower-secondary education	−0.0004	0.0000	−0.1062***	−0.0580***	−0.0555***	0.0160***	0.0023	−0.0203***
	(0.0010)	(0.0001)	(0.0077)	(0.0138)	(0.0133)	(0.0035)	(0.0153)	(0.0034)
Upper-secondary education	0.0041***	0.0000	−0.0386***	0.1000***	0.0906***	0.0429***	0.0581***	−0.0255***
	(0.0012)	(0.0002)	(0.0091)	(0.0139)	(0.0134)	(0.0044)	(0.0161)	(0.0035)
Post-secondary education	0.0101***	−0.0005***	0.2126***	0.4732***	0.4774***	−0.0057	0.2906***	−0.0419***
	(0.0019)	(0.0002)	(0.0167)	(0.0207)	(0.0192)	(0.0071)	(0.0181)	(0.0044)
Urban area (urban = 1, rural = 0)	0.0098***	0.0006***	0.0386***	0.0375***	0.0337***	0.0665***	0.0443***	0.0027
	(0.0009)	(0.0002)	(0.0072)	(0.0064)	(0.0061)	(0.0053)	(0.0068)	(0.0029)
Interaction between explanatory variables and sample 2019–2020	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample-pair dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	0.1595***	0.0010***	0.6331***	0.3483***	0.2314***	3.2107***	7.4784***	0.3940***
	(0.0064)	(0.0003)	(0.0252)	(0.0268)	(0.0235)	(0.0218)	(0.0289)	(0.0154)
Observations	2,045,740	2,045,871	2,045,871	2,045,871	2,045,871	1,986,745	869,660	869,660
R-squared	0.032	0.017	0.247	0.348	0.343	0.219	0.354	0.091
Robust standard errors in parentheses * p < 0.01, p < 0.05, * p < 0.1

Explanatory variables	Groups of workers with different wage levels
	Quintile 1	Quintile 2	Quintile 3	Quintile 4	Quintile 5	Bottom 10 %	Bottom 40 %	Below minimum wages
Quarter 2–4 * COVID year	−0.1350***	−0.0139***	−0.0099	−0.0108	−0.0150	−0.1273***	−0.1413***	−0.1937**
	(0.0385)	(0.0032)	(0.0060)	(0.0097)	(0.0199)	(0.0339)	(0.0350)	−0.0702
Quarter 2–4	0.0329***	−0.0018	−0.0072***	−0.0058*	−0.0759***	0.0250***	0.0323***	−0.0049
	(0.0034)	(0.0029)	(0.0015)	(0.0033)	(0.0064)	(0.0057)	(0.0031)	(0.0058)
COVID year	0.1022***	−0.0217***	−0.0003	0.0320***	0.0200***	0.0370***	0.1709***	0.2181***
	(0.0072)	(0.0028)	(0.0014)	(0.0033)	(0.0056)	(0.0086)	(0.0069)	(0.0099)
Male (male = 1, female = 0)	0.0739***	0.0063***	0.0026***	0.0080***	0.0501***	0.0526***	0.0977***	−0.0118
	(0.0047)	(0.0010)	(0.0005)	(0.0007)	(0.0045)	(0.0058)	(0.0047)	(0.0079)
Age	0.0339***	0.0018***	0.0008***	0.0013***	0.0072***	0.0284***	0.0405***	0.0121***
	(0.0011)	(0.0001)	(0.0001)	(0.0002)	(0.0016)	(0.0012)	(0.0013)	(0.0014)
Age squared	−0.0005***	−0.0000***	−0.0000***	−0.0000***	−0.0001***	−0.0004***	−0.0006***	−0.0002***
	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)	(0.0000)
Less than primary education	Reference
Primary education	0.0589***	−0.0010	−0.0008	−0.0121***	−0.0573***	0.0486***	0.0532***	0.0314***
	(0.0078)	(0.0019)	(0.0010)	(0.0022)	(0.0163)	(0.0058)	(0.0127)	(0.0076)
Lower-secondary education	0.0802***	0.0019	−0.0006	−0.0104***	−0.0533***	0.0591***	0.0837***	0.0370***
	(0.0092)	(0.0022)	(0.0012)	(0.0027)	(0.0150)	(0.0063)	(0.0142)	(0.0083)
Upper-secondary education	0.0622***	0.0047**	0.0014	−0.0026	−0.0257	0.0295***	0.0844***	0.0086
	(0.0111)	(0.0022)	(0.0014)	(0.0029)	(0.0168)	(0.0083)	(0.0163)	(0.0098)
Post-secondary education	0.0982***	0.0110***	0.0052***	0.0109***	0.0664***	0.0631***	0.1443***	−0.0097
	(0.0133)	(0.0024)	(0.0013)	(0.0030)	(0.0182)	(0.0113)	(0.0186)	(0.0161)
Urban area (urban = 1, rural = 0)	0.0227***	0.0034***	0.0008	0.0015	−0.0042	0.0257***	0.0239***	0.0170**
	(0.0051)	(0.0008)	(0.0006)	(0.0011)	(0.0046)	(0.0057)	(0.0051)	(0.0077)
Year fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
District fixed-effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Constant	7.4627***	8.5720***	8.7577***	8.9119***	9.2160***	7.4063***	7.4470***	7.5715***
	(0.0257)	(0.0044)	(0.0021)	(0.0062)	(0.0417)	(0.0235)	(0.0325)	(0.0283)
Observations	416,405	221,476	158,396	185,865	142,605	240,272	637,881	50,203
R-squared	0.183	0.070	0.086	0.068	0.105	0.156	0.222	0.344
Note: Robust standard errors in parentheses (corrected for sampling weight and clustered at the district level and year-by-quarter level). * p < 0.01, p < 0.05, * p < 0.1. Source: authors’ estimations from the LFSs.

Explanatory variables	Industries of employment
	Agriculture	Fishery	Mining, electricity, water	Manufacture, processing	Construction	Trade	Hotel, restaurant	Transportation	Service
Effects on workers below minimum wages	−0.0403	−0.1101*	−0.0331	−0.1138***	−0.1442***	−0.3687*	−0.2739**	−0.2463***	−0.0971***
Effects on workers below minimum wages	(0.0357)	(0.0534)	(0.0329)	(0.0248)	(0.0309)	(0.1961)	(0.1184)	(0.0588)	(0.0294)
Effects on workers not below minimum wages	0.0219	−0.0125	0.0867***	0.0011	−0.0188*	0.0576	0.0565	−0.0361	−0.0194***
Effects on workers not below minimum wages	(0.0189)	(0.0327)	(0.0110)	(0.0236)	(0.0093)	(0.0695)	(0.0751)	(0.0314)	(0.0066)
Effects on workers in wage quintile 1	−0.0312	−0.2405***	−0.0532	−0.0596	−0.1273***	−0.5564***	−0.3454***	−0.3456***	−0.1020***
Effects on workers in wage quintile 1	(0.0417)	(0.0597)	(0.1249)	(0.0353)	(0.0336)	(0.1326)	(0.1166)	(0.0680)	(0.0219)
Effects on workers in wage quintiles 2 to 5	−0.0289***	0.0036	0.0415**	−0.0170	−0.0390***	0.0713	0.0361	−0.0693***	−0.0122
Effects on workers in wage quintiles 2 to 5	(0.0066)	(0.0228)	(0.0153)	(0.0239)	(0.0057)	(0.0684)	(0.0802)	(0.0218)	(0.0084)
Note: Control variables are the same as in Table 1, including gender, age, age squared, education levels and urban dummy, district and year fixed-effects.Robust standard errors in parentheses (corrected for sampling weight and clustered at the district level and year-by-quarter level) . * p < 0.01, p < 0.05, * p < 0.1. Source: authors’ estimations from the LFSs.

Explanatory variables	Areas and regions
	Rural	Urban	Red River Delta	Northern Mountain	Central Coast	Highland	Southeast	Mekong River Delta
Effects on workers below minimum wages	−0.1138***	−0.1759***	−0.0799***	−0.0681***	−0.0950***	−0.0246	−0.1176***	−0.0644***
Effects on workers below minimum wages	(0.0289)	(0.0584)	(0.0214)	(0.0095)	(0.0144)	(0.0228)	(0.0149)	(0.0076)
Effects on workers not below minimum wages	0.0164	−0.0257	0.0179	0.0103	0.0206	−0.0086	−0.0349	−0.0007
Effects on workers not below minimum wages	(0.0238)	(0.0158)	(0.0230)	(0.0112)	(0.0144)	(0.0108)	(0.0227)	(0.0218)
Effects on workers in wage quintile 1	−0.1464**	−0.2415***	−0.1960**	−0.2438***	−0.2558***	−0.0040	−0.1025	−0.1320*
Effects on workers in wage quintile 1	(0.0655)	(0.0743)	(0.0934)	(0.0531)	(0.0460)	(0.0577)	(0.0938)	(0.0712)
Effects on workers in wage quintiles 2 to 5	−0.0066	−0.0309*	0.0040	−0.0049	−0.0036	−0.0295***	−0.0486**	−0.0196
Effects on workers in wage quintiles 2 to 5	(0.0208)	(0.0157)	(0.0232)	(0.0204)	(0.0077)	(0.0086)	(0.0212)	(0.0214)
Note: Control variables are the same as in Table 1, including gender, age, age squared, education levels and urban dummy, district and year fixed-effects.Robust standard errors in parentheses (corrected for sampling weight and clustered at the district level and year-by-quarter level) . * p < 0.01, p < 0.05, * p < 0.1. Source: authors’ estimations from the LFSs.

PERMALINK

Did a successful fight against COVID-19 come at a cost? Impacts of the pandemic on employment outcomes in Vietnam

Hai-Anh H Dang

Cuong Viet Nguyen

Calogero Carletto

Abstract

1. Introduction

Fig. 1.

2. Country background and data description

2.1. Country background

2.2. Data description

Fig. A3.

3. Estimation method

4. Empirical results

4.1. Testing assumptions

Fig. A2.

4.2. Estimated pandemic impacts

Table 1.

Fig. 2.

4.3. Further analysis of potential mechanism

Table 2.

Table 3.

4.4. Robustness checks

4.5. Heterogeneity analysis

Fig. 3.

Fig. A5.

Fig. A6.

5. Effects on low-wage workers and wage inequality

Fig. 4.

Table 4.

6. Discussion and conclusion

CRediT authorship contribution statement

Declaration of Competing Interest

Acknowledgements

Footnotes

Appendix A.

Fig. A1.

Fig. A4.

Fig. A7.

Table A1.

Table A2.

Table A3.

Table A4.

Table A5.

Table A6.

Table A7.

Table A8.

Table A9.

Table A10.

Table A11.

Table A12.

Table A13.

Table A14.

Table A15.

Table A16.

Table A17.

Table A18.

Table A19.

Table A20.

Table A21.

Table A22.

Table A23.

Table A24.

Table A25.

Table A26.

Table A27.

Table A28.

Table A29.

Table A30.

Table A31.

Table A32.

Table A33.

Table A34.

Table A35.

Table A36.

Data availability

References

Associated Data

Data Availability Statement

ACTIONS