Skip to content
Sign inGet started free

The record · 8 September 2026

In Ethiopia, 340 people work alone for every one who employs anybody.Being self-employed rarely means employing anyone.

Fourteen economies, one year, one classification. "Self-employed" is an aggregate of three statuses that behave nothing like each other, and separating them changes what a labour market looks like. Everything below is checkable against the live market.

· 8 min read

54.5%of everyone working in Ethiopia is an own-account worker, self-employed with nobody working for them
0.16%of the same total are employers, the lowest share of the 182 economies reporting it, and 340 times smaller than the figure beside it
6.0own-account workers per employer at the middle of the distribution, across all 182 economies; Brazil is the 91st of the 182 and sits there

Two shares of the same total

For each economy there are two published shares of the same denominator, everyone in work: the share who are own-account workers and the share who are employers. Because both are shares of the same total, dividing one by the other gives a count, and the count is how many people work alone for each person who employs anybody.

At the top of the table the two barely belong to the same economy. Ethiopia records 54.5% own-account and 0.16% employers, which is 340 people working alone per employer. Moldova records 56.1% and 0.18%, or 319. Mali runs 48.3% against 0.60%, and Laos, not in the table, 54.2% against 0.76%. Only three of the 182 economies pass 100.

The table then descends through the markets this desk publishes in. Viet Nam 39.4% against 1.9%, India 55.9% against 3.3%, Thailand 33.2% against 2.7%, Indonesia 37.0% against 3.4%, the Philippines 27.4% against 2.5%. The median across all 182 is 6.0, and Brazil is the 91st of the 182, sitting exactly at that middle. At the foot Germany records 1.1 and Qatar 0.3, which is to say Qatar has three employers for every ten people working alone.

Provenance: Source
Own-account workers and employers as shares of total employment, 2024, for fourteen economies. Sorted by the count of own-account workers per employer, highest first. Both columns are shares of the same total, everyone in work, which is what makes the third column a legitimate division.
EconomyOwn-accountEmployersPer employer
Ethiopia54.5%0.16%339.5
Moldova56.1%0.18%319.4
Mali48.3%0.60%80.7
Viet Nam39.4%1.95%20.2
India55.9%3.33%16.8
Thailand33.2%2.69%12.4
Indonesia37.0%3.37%11.0
Philippines27.4%2.53%10.8
United Kingdom11.3%1.61%7.0
Brazil25.1%4.18%6.0
Türkiye16.2%4.63%3.5
Nigeria56.7%19.68%2.9
Germany4.1%3.81%1.1
Qatar0.1%0.38%0.3

Why the ratio moves

The classification is the mechanism. "Self-employed" in the World Bank's aggregate means employers plus own-account workers plus contributing family workers, and the three are not variations on a theme: one hires, one works alone, one is unpaid in a family business. A country can be 85% self-employed and contain almost nobody who employs anyone, because the aggregate never promised otherwise.

What separates the top of this table from the bottom is where wage employment is scarce. Ethiopia records 15.4% of employment as wage and salaried and Qatar 99.2%. Where there are few employers there are few jobs to hold, and a person who needs an income works alone, which raises the own-account share and the ratio at the same time. That is one relationship read two ways, not two findings.

It also explains the piece's most useful refusal. Nigeria records 19.7% of employment as employers, the second highest share on earth after Djibouti, and still comes out at 2.9 because its own-account share is 56.7%. Plenty of people employing somebody, and even more people working alone. The ratio is a shape, and two economies with the same shape can be nothing alike in size.

Provenance: SourceWhat the classification asks
  1. Does the job pay a wage or salary agreed with someone else? Then the worker is wage and salaried, and none of this page is about them.
  2. Is the job self-employed, with one or more employees engaged on a continuous basis? Then the worker is an employer. One employee is enough; there is no threshold of size, turnover or formality.
  3. Is the job self-employed with no such employees? Then the worker is own-account, unless the work is unpaid in a family business, which makes them a contributing family worker instead.

Nothing in those questions asks whether the work is registered, taxed, ambitious or secure. That is why the ratio cannot be read as a measure of informality or of enterprise: a licensed accountant working alone and a woman selling tomatoes at the roadside are both own-account, and a market trader with one paid assistant is an employer while a start-up founder with no staff is not.

Status in employment is decided by three questions, in the international classification the ILO uses and the World Bank publishes against. They are questions about the job, not about the person or the size of the business.

What the ratio does not say

It is not a ranking of rich against poor, and two rows make that impossible to miss. The United Kingdom sits at 7.0, above Brazil at 6.0 and China at 6.0, with 78 of the 182 economies above it. And Moldova, upper-middle-income and European, is second in the world. A reader who takes the top of this table as a poverty ranking has to explain both.

It is not a measure of entrepreneurship, in either direction. Nothing in the classification asks whether a business was founded, registered, funded or intended to grow, and a person is an employer the moment one other person works for them continuously. Neither a high nor a low ratio tells you how many businesses were started.

It is not a measure of informality either. Informality is defined by whether the job is covered by employment law and social protection, which is a different question with its own indicators, and a licensed professional working alone is own-account in exactly the same cell as an unregistered street trader.

What the data establishes

  • In 2024 own-account workers were 54.5% of employment in Ethiopia and employers 0.16%, a count of 340 own-account workers per employer, the highest of the 182 economies reporting all four series.
  • Moldova is second at 319, with 56.1% own-account against 0.18% employers, and is an upper-middle-income European economy.
  • The median across those 182 economies is 6.0 own-account workers per employer, and Brazil, the 91st of the 182, sits at that middle. Three economies exceed 100 and eight record more employers than own-account workers.
  • Nigeria records the second highest employer share on earth at 19.7% of employment, after Djibouti at 20.9%, and still comes out at 2.9 because 56.7% of its employment is own-account.

What it does not

  • That a high count measures entrepreneurship, or that a low one measures its absence. The classification never asks whether a business was founded, registered or meant to grow.
  • That it measures informality. Informality is defined by legal and social-protection coverage, has its own indicators, and does not separate a licensed professional working alone from an unregistered trader.
  • That it ranks rich economies against poor ones. The United Kingdom is at 7.0 with 78 economies above it, above Brazil and China at 6.0, and Moldova is second in the world.
  • Anything about firms. Every figure is a share of PEOPLE IN WORK; none of them counts businesses, and no number here divides by a company.
  • Any cause. Fourteen economies show a relationship in one year, and nothing here says which way any influence runs or what would happen if it changed.
  • Anything about the 35 economies in the World Bank's country list that do not carry all four series for 2024, or about 2025 and 2026. The latest year carrying all four is 2024.

Method

  1. One source, one year, four series. World Bank Open Data SL.EMP.MPYR.ZS (employers), SL.EMP.VULN.ZS (vulnerable employment), SL.FAM.WORK.ZS (contributing family workers) and SL.EMP.SELF.ZS (self-employed, total), each as a share of total employment and each a modelled ILO estimate. Reference year 2024, read on 8 September 2026, dataset last updated 2026-07-13 by the source's own stamp.
  2. ONE COLUMN IS DERIVED AND SAYS SO. The Bank publishes no own-account series for 2024, so own-account is computed as vulnerable employment minus contributing family workers, which is the identity the classification defines. It was checked against the published aggregate on all 182 economies: self-employed minus employers equals vulnerable employment to within 0.0023 percentage points, and self-employed plus wage and salaried equals 100 to within 0.0009. The per-employer count is computed from unrounded values and rounded once for display.
  3. AGGREGATES ARE NOT COUNTRIES, and they carry three-letter codes too. "World", "Sub-Saharan Africa" and every income group were removed using the World Bank's own flag for them, which is what leaves 182 economies carrying all four series for 2024 out of the 217 in its country list. A first pass that filtered on code length alone let 41 aggregates in, and every count and superlative here was recomputed after they were removed.
  4. These are MODELLED estimates, not survey readings. The ILO fits them where national labour force surveys are absent or not comparable, and for several economies here the 2024 value is imputed rather than observed. Ethiopia and Moldova were checked back to 2014 for a level break and have neither: Ethiopia's employer share drifts from 0.45% to 0.16% across the decade and Moldova's from 0.41% to 0.18%, both without a step.
  5. The fourteen economies span the range from the highest count to the lowest and cover eleven of the sixteen languages this edition ships in. Every economy carrying all four series for 2024 was eligible; nothing was excluded for being inconvenient, and the whole 182 stands behind every count in the prose.

BabZituna scores every job against your profile on six real dimensions, and publishes what it can and cannot measure. See the live market.

The markets in this piece, live

Earlier editions