Survey microdata · Labour markets · 2007–2015
What kinds of jobs do Vietnam’s small and medium-sized manufacturers offer? How did wages, contracts and social insurance change as the economy grew, how do workers find these jobs, and does training pay? An analysis of five waves of a worker survey, 2007–2015.
The survey interviews a few workers in each firm, 2.4 on average, in ten provinces. In every wave about 41% of the workers are women, and between half and three in five are production workers. The workforce is ageing slowly: the average worker was 34 in 2007 and 37 in 2015, and average time with the current firm rose from 6.3 to 7.7 years.
Education rose. The share of workers with primary schooling or less halved, from 10.8% to 5.1%, while the share with vocational or technical training rose from 22% to 28% and the share with a college or university degree from 18% to 22%.
Percent of workers interviewed in each survey year.
Family ties are common. In 2011–2015, about a third of the workers interviewed (32–33%) were the owner, a member of the owner’s household, or another relative of the owner.
Among workers paid by the month, about four in five, the median monthly wage rose from 1.35 million dong in 2007 to 4.0 million dong in 2015. Consumer prices roughly doubled over the same period, so in 2015 prices the median rose from 2.8 to 4.0 million dong: an increase of 43%, or about 4.6% a year. Men’s median real wage rose from 2.90 to 4.50 million dong and women’s from 2.70 to 4.00 million.
Thousand dong per month in 2015 prices; workers paid by the month.
Women earn less than comparable men. Comparing workers with the same education, age, tenure, job type and province, women earned between 86% and 91% of men’s wages in every wave, and the gap changed little over the period. The unadjusted gap is smaller because women in these firms are better educated than men on average: 26% of women interviewed had a college or university degree, against 13% of men.
Unadjusted: women’s median wage divided by men’s. Adjusted: from a regression of log wages on sex, education, age, tenure, job type and province, estimated separately for each wave.
Education pays. Holding sex, age, tenure, province and survey year constant, workers with a college or university degree earned 55% more than workers with primary schooling or less, workers with vocational or technical training 28% more, and workers who finished upper secondary school 19% more.
Percent difference in monthly wage relative to workers with primary schooling or less, holding sex, age, tenure, province and year constant; 2007–2015 pooled.
Just under half of workers had a formal labour contract in 2009 (47.9%); by 2015 the share had reached 54.0%. Employer-paid benefits spread faster. The share of workers whose employer paid for health insurance rose from 26% in 2007 to 48% in 2015, paid annual leave from 28% to 48%, and paid sick leave from 33% to 53%.
Vietnam launched unemployment insurance in 2009 and widened it from 2015 under its Employment Law, which extended coverage to employers of any size and to contracts of three months or more. In the survey, the share of workers whose employer paid into unemployment insurance rose from 28% in 2011 to 41% in 2015. Coverage is almost entirely tied to the labour contract: across 2011–2015, 64% of workers with a formal contract had unemployment insurance, against 3% of workers without one.
Percent of workers. The contract question was first asked in 2009 and the unemployment-insurance question in 2011.
The gaps between groups are large. In 2015, 89% of workers with a college or university degree had a formal contract, against 37% of those with upper secondary schooling or less. Only 32% of the owner and owner’s relatives had one, against 64% of workers with no family tie. Women were more likely than men to have a contract (60% against 50%) and unemployment insurance (46% against 38%).
Percent of workers in each group with a formal labour contract and with employer-paid unemployment insurance.
Between 2009 and 2015, 72% of workers said they found their job through family or friends: 29–35% through the owner or a relative working in the firm, and 39–40% through a friend or acquaintance. Only 9% found it through an employment agency or a job advertisement, although that share rose from 6.2% in 2011 to 11.6% in 2015.
Percent of workers in each survey year. “Family” is the owner as a relative or a relative working in the firm; “Friends” is the owner as a friend or a friend working in the firm.
Formal channels matter mainly for educated workers: 15% of college or university graduates found their job through an agency or advertisement, against 3–4% of workers with lower secondary schooling or less.
Percent of workers; 2009–2015 pooled.
Why it matters. When most hiring runs through personal networks, workers without contacts inside a firm hear about fewer vacancies, and public employment services see only a small part of the market. Measuring how jobs are actually found is a starting point for designing job-matching services that reach those workers.
In 2007–2011, between 31% and 40% of workers said they had received general or on-the-job training at their current firm. From 2013 the survey asked separately about formal on-the-job, off-the-job and informal training, so the two periods are not directly comparable. In 2015, 22% of workers had received formal on-the-job training, 2.5% off-the-job training and 37% informal training; 44% had received at least one kind.
Percent of workers. The question changed in 2013, so the series are shown separately.
Training goes to workers who already have more. In 2007–2011, 40% of college or university graduates had been trained at their firm, against 26–27% of workers with lower secondary schooling or less; in 2009–2011, 44% of workers with a formal contract had been trained, against 28% of those without one.
Percent of workers who received general or on-the-job training at their current firm.
Does training raise wages? In 2007–2011, trained workers earned about 11% more than untrained workers. But because training goes to more educated workers in better firms, that raw gap mixes the effect of training with the effect of who is trained and where. Figure 11 makes the comparison progressively fairer. Comparing workers with the same education, age, tenure, sex, job type and province lowers the premium to 7%; comparing trained workers with untrained co-workers in the same firm lowers it to 6%. In each wave on its own, the same-firm estimate is well below the raw gap. Following the same workers from 2013 to 2015, those who started training in between saw their wages grow about 5% more than others, but with few such workers the estimate is imprecise: the 95% interval runs from −7% to +17%.
Percent difference in monthly wage; dots are estimates and lines are 95% confidence intervals, with standard errors clustered by firm. The last row uses the 2013–2015 training questions and workers interviewed in both waves.
Workers’ own reports point the same way: in 2013–2015, 45% of trained workers said their wage rose after the training.
What this means for evaluating training. A simple comparison of trained and untrained workers overstates what training delivers, because the two groups differ before any training takes place. A credible evaluation of a training program needs a comparison group that resembles the participants — through random assignment, matching, or following the same people over time — and enough participants to measure the effect precisely.
Because the survey returns to the same firms, it can follow workers over time. Among firms surveyed in consecutive waves, 43% of the workers interviewed in 2011 or 2013 were interviewed again in the same firm two years later. Younger workers move more: 37% of workers under 30 were still there, against 47% of those aged 30 or older. Having a formal contract or training made little difference. Workers who stayed saw their real wage rise by a median of 10.7% over the two years, or about 5% a year.
Percent of workers interviewed in 2011 or 2013 in firms surveyed again two years later. Because only a few workers are interviewed per firm, this understates how many workers stay; differences between groups are more informative than the level.
Job quality varies widely across the ten provinces (2011–2015 pooled). The share of workers with a formal contract ranges from 20% in Phu Tho to 71% in Ha Noi city, and unemployment-insurance coverage from 13–14% in Phu Tho and the former Ha Tay province to 53% in Ho Chi Minh City. Median real wages are highest in Ha Noi and Hai Phong (4.0 million dong) and lowest in Khanh Hoa (3.0 million). Hiring through family and friends is most common in Nghe An and Hai Phong (85% and 84%) and least common in Lam Dong (45%).
Choose an indicator. Darker provinces have higher values; grey provinces are not in the survey. Ha Tay merged into Ha Noi in 2008, so the map shows Ha Noi including former Ha Tay, while the ranking shows both. Hover over a province or a bar for details.
FP.CPI.TOTL) and expressed in 2015 prices.