Bangladesh's jobs are drifting back to the farm — and young women are the most shut out

Introduction

For a quarter of a century, Bangladesh’s economic story was one of workers leaving the field for the factory and the shop. Between 1991 and 2017, the share of employed people working in agriculture fell from 69.9% to 40.6%. By 2025 it had climbed back to 44.3%.¹

40.6% → 44.3%

Share of Bangladesh’s employed people working in agriculture, 2017 and 2025 (modelled ILO estimates).

At the same time, the country’s measured unemployment rate looks low — 3.8% in 2025 — while the rate for people aged 15 to 24 is 9.4%.² Neither number captures the full picture. This article looks at where Bangladesh’s jobs are, who is being left out, and what the data say about the quality of the work on offer.

All the employment figures here are International Labour Organization (ILO) estimates, published through the World Bank. They are modelled from labour force surveys, so small year-to-year moves should be read with caution; the multi-year direction is more reliable.

The structural shift has stalled

Agriculture’s share of employment fell from 69.9% in 1991 to 47.6% in 2010 and bottomed out at 40.6% in 2017.¹ Industry’s share peaked at 20.8% in 2013, and services reached 39.4% in 2019.³⁴

Since then the movement has run the other way. By 2025 agriculture employed 44.3% of workers, industry 18.0% and services 37.7%. Compared with their peaks, the industry share is down by 2.8 percentage points and the services share by 1.7.¹³⁴

The data do not say why. The shift coincides with the COVID-19 years, when agriculture is often a fallback for workers who lose other work, but the series shows no recovery of industry or services afterwards. That is a question for the national labour force surveys, not for modelled estimates.

What the numbers do establish is that the long process of moving people into higher-productivity work is no longer visibly advancing. For a country with a large, young workforce, that matters.

Youth unemployment: higher, and unequal by sex

The unemployment rate among people aged 15 to 24 was 9.4% in 2025, about two and a half times the 3.8% rate for all ages.²⁵ It was 6.3% in 2010 and has been above 10% in most years since 2016, peaking at 13.5% in 2020 and falling to 7.9% in 2023 before rising again.²

The split by sex is striking. In 2025, 13.3% of young men in the labour force were unemployed, against 5.3% of young women.⁶⁷ Young women’s rate has fallen sharply, from a peak of 16.2% in 2017 to 4.7% in 2023.⁷

A lower unemployment rate for women is not necessarily good news. The rate counts only people who are looking for work. A woman who has stopped looking, or never started, is outside the labour force and does not appear in it. Across all women aged 15 and over, labour force participation was 38.7% in 2025, down from 43.7% in 2022, against about 80% for men.⁸

Outside both school and work

A better measure of exclusion is the share of young people who are not in education, employment or training (NEET). In 2024 it was 19.7% for young women and 10.9% for young men, based on ILO labour force statistics.⁹¹⁰

19.7% vs 10.9%

Young women and young men aged 15–24 not in education, employment or training, 2024.

Young women are therefore almost twice as likely as young men to be doing none of the three. Their low unemployment rate and high NEET rate together suggest that many are outside the labour market rather than successfully inside it. The NEET series has gaps — survey years are missing between 2017 and 2022 — and earlier values come from different survey rounds, so this article does not draw a trend from them.

Work that is insecure

Even for those with jobs, quality is a concern. Vulnerable employment — own-account workers and contributing family workers, who typically lack formal contracts and social protection — made up 57.3% of all employment in 2025.¹¹ It fell steadily from 68.5% in 1991 to 56.2% in 2017, then edged back up to a range of 57–58% in every year since 2020.¹¹

Taken together, the data describe a labour market that grew in the 1990s and 2000s, but is not currently creating enough formal, non-farm jobs for new entrants, particularly women.

What it will take

The data point to the problem rather than the cure, so these options are framed as priorities for investigation and policy, not as proven fixes:

• Better, more frequent labour data. Modelled estimates and gaps in the NEET series limit what can be said with confidence. Regular, comparable labour force surveys with youth and sex breakdowns would let policymakers see whether the farm drift is real and lasting.
• Non-farm job creation. With industry and services both below their earlier peaks in employment share, expanding productive jobs outside agriculture is the central test for the next decade.
• Pathways for young women. A NEET rate of nearly one in five suggests barriers that unemployment statistics do not show, such as safety, transport, childcare and norms. Programmes should be evaluated against NEET and participation rates, not unemployment alone.
• Skills matched to demand. Training for young people who have left school should be tied to employers’ actual needs and tracked for job outcomes.
• Protection for informal workers. With more than half of workers in vulnerable employment, extending social protection to own-account and family workers is as important as creating new jobs.