Skip to content
Sign inGet started free
Photo: Guilherme Ramos / Unsplash · Argentina

Market watch · 57

Argentina’s cities have more workers and fewer employees

The number of people who employ others rose over the same three years.

Argentina’s household survey sorts everybody in work by the status of their own job: working for an employer, working for yourself with staff, working for yourself without staff, or working unpaid in a family member’s business. All four count as employment, which is why the total can rise in the same three years in which the number of people with an employer falls. Here is what each category counts, which of them absorbed the whole of the increase, and what it means if you are hiring or looking for work in Argentina’s cities.

+434,363more people in work in the urban areas Argentina’s survey covers than three years earlier
-169,851fewer of them working for an employer over the same three years
70.6%of everybody in work is an employee, against 74.2% three years earlier

Seven in ten of Argentina’s urban workers have an employer

Employees9,554,718
Own-account3,447,070
Employers470,627
Family workers69,141
Employment by status in employment, the urban agglomerations covered by Argentina’s permanent household survey, second quarter of 2026, people. The four categories shown are the whole of employment in this quarter: the classification’s sixth category, workers not classifiable by status, is not published for it, so the four add to the 13,541,556 total exactly. An own-account worker and an employer both work for themselves, and the difference between them is whether anybody else is employed. The bars are levels, not changes, because one of the changes this edition is about is negative and a bar cannot start a negative value at zero. ILO, employment by sex and status in employment, indicator EMP_TEMP_SEX_STE_NB_Q, read 10 November 2026.

Two kinds of work, two directions

In the second quarter of 2026, 13,541,556 people were in work in the urban areas Argentina’s household survey covers, against 13,107,193 three years earlier. That is 434,363 more, a rise of 3.3% on this edition’s subtraction of the two published counts.

Over the same three years the number of them working for an employer fell, from 9,724,569 to 9,554,718. That is 169,851 fewer, a fall of 1.7%, in a period when the total rose.

Both statements are true because all four categories count as employment. Own-account workers, who work for themselves and employ nobody, rose by 522,761 to 3,447,070. People who employ at least one other person rose by 55,116 to 470,627. People working unpaid in a family member’s business rose by 26,337 to 69,141. Those four changes add to 434,363, the change in the total, exactly.

Provenance: SourceWhat the survey counts
  1. AN EMPLOYEE works for an employer and is paid by that employer. In the data this is both the ICSE-93 category 1 and the aggregate called Employees, and the two are identical in all 177 comparisons, which is the control that fixes the mapping of code to category.
  2. AN OWN-ACCOUNT WORKER works for themselves and employs nobody. AN EMPLOYER works for themselves AND has at least one employee. That is the whole difference between those two, and both of them are working for themselves rather than for anybody else.
  3. A CONTRIBUTING FAMILY WORKER works in a family member’s business WITHOUT PAY and is counted as employed all the same. There are 69,141 of them here, one in 196 of everybody in work, and the absence of pay is the definition of the category rather than anything this read measured.
  4. ALL FOUR COUNT AS EMPLOYMENT. That is the whole reason the total can rise while the number of employees falls, and the two statements are not in conflict. The data also carries an aggregate called Self-employed, which equals employers plus own-account workers plus contributing family workers exactly, every quarter.
  5. AN EMPLOYER IS A PERSON, NOT A FIRM. The survey asks people what status their own job has. 470,627 is a number of people who employ somebody; it is not a number of businesses, and nothing here counts enterprises, payrolls, jobs or vacancies. More employers beside fewer employees therefore says nothing at all about how many people any business employs: that was not read.
  6. AND THE PLACE IS NOT THE COUNTRY. The data labels its own coverage "Main cities or metropolitan areas". Every count here is of the urban areas the survey visits, never of Argentina as a whole.

So these counts say how many people held each kind of working arrangement in one quarter, in the urban areas the survey covers, and nothing about why anybody held it, what it paid, or how long it lasted.

Four categories and one aggregate, all named by the data itself. A reading that merges an employee with somebody working on their own account, or that treats an employer as a business, says what the data do not.

Three years, four categories

Employees are seven in ten of everybody in work, 9,554,718 of 13,541,556, or 70.6%. Own-account workers are 25.5%, employers 3.5% and contributing family workers 0.5%. Three years earlier employees were 74.2%, so their share has fallen 3.6 percentage points.

Of the four categories, employees are the only one whose count fell. Own-account workers grew 17.9%, employers 13.3% and contributing family workers 61.5%, the last on a base of 42,804 and therefore the least steady figure in the table.

The four categories are the whole of employment in both quarters. The classification has a sixth category, workers not classifiable by status, and it is published for neither of these quarters, so there is no residual here and the rows add to the totals exactly rather than approximately.

Provenance: Source
Employment by status in employment, the urban agglomerations covered by Argentina’s permanent household survey, people, second quarter of 2023 against second quarter of 2026: the same quarter three years apart, because a second quarter and a fourth quarter are not alike. The change column is this edition’s subtraction of the two published counts in the same row. The four categories add to the total exactly in both quarters, because the classification’s sixth category, workers not classifiable by status, is published for neither. ILO, indicator EMP_TEMP_SEX_STE_NB_Q, ICSE-93.
Status in employmentSecond quarter 2023Second quarter 2026Change
Employees, working for an employer and paid by that employer9,724,5699,554,718-169,851
Own-account workers, working for themselves with no employees2,924,3093,447,070+522,761
Employers, working for themselves with at least one employee415,511470,627+55,116
Contributing family workers, unpaid in a family member’s business42,80469,141+26,337
All people in work13,107,19313,541,556+434,363

One year, two years, three years

The pattern does not depend on which window is chosen. Over one year employees fell 74,248 while the total rose 239,579. Over two years employees fell 249,454 while the total rose 256,143. Over three years employees fell 169,851 while the total rose 434,363. Every window ends in the second quarter of 2026 and every one begins in a second quarter.

Own-account workers rose by more than the total in all three windows, by 289,719 over one year, 391,629 over two and 522,761 over three, which is the arithmetic consequence of the only falling category being the biggest one.

The two-year window holds the largest fall in employees, and the one-year window the smallest rise in the total, and both are printed here rather than left out. A reader who wants the least favourable reading of this finding should take the one-year window, where employees fell 74,248 and the people who employ others rose by just 3,116.

Provenance: Source
The same comparison over one, two and three years, every window ending in the second quarter of 2026 and every one of them starting in a second quarter. This table is here because one window could be a window picked to suit the finding, and the answer does not depend on the choice: in all three, employees fall, own-account workers rise by more than the total, and the total rises. Each row is this edition’s subtraction of two published counts, and in each row the four category changes add to the change in the total exactly. ILO, indicator EMP_TEMP_SEX_STE_NB_Q.
Window ending second quarter 2026EmployeesOwn-account workersAll people in work
One year, from the second quarter of 2025-74,248+289,719+239,579
Two years, from the second quarter of 2024-249,454+391,629+256,143
Three years, from the second quarter of 2023-169,851+522,761+434,363

Thirty-six for every hundred

There were 36.08 own-account workers for every 100 employees in the second quarter of 2026, against 30.07 three years earlier and 29.35 in the second quarter of 2019. The ratio is this edition’s arithmetic on two published counts rather than a figure the ILO prints.

That is not the highest the series holds, and this edition does not say it is. The fourth quarter of 2020 holds 37.40 per 100, and the same quarter holds the lowest employee share, 70.53% against 70.56% now. Of the forty-one quarters from the first quarter of 2016, the latest reading is the second highest ratio and the second lowest employee share, not the first of either.

What is at its highest is the number of own-account workers itself, 3,447,070, the largest of those forty-one quarters. The number of employees is not at a low: 9,554,718 is the fifteenth highest of the forty-one, and the most the series holds is 10,023,469 in the third quarter of 2023.

Provenance: Source
Employees and own-account workers in six quarters, the urban agglomerations covered by Argentina’s permanent household survey, people, with own-account workers per 100 employees. The ratio is this edition’s arithmetic on the two published counts in the same row and is not a figure the ILO prints. The fourth quarter of 2020 is in this table although every other row is a second quarter: it holds 37.40 per 100, the highest ratio of the forty-one quarters read, and the lowest employee share. That is why nothing here calls the latest reading a record. ILO, indicator EMP_TEMP_SEX_STE_NB_Q.
QuarterEmployeesOwn-account workersOwn-account per 100 employees
Second quarter of 20198,944,2902,625,53229.35
Fourth quarter of 20208,122,9823,038,16637.40
Second quarter of 20239,724,5692,924,30930.07
Second quarter of 20249,804,1723,055,44131.16
Second quarter of 20259,628,9663,157,35132.79
Second quarter of 20269,554,7183,447,07036.08

What these figures do not say

They do not say why anybody is counted where they are. The survey records the status of a person’s job, never the reason for it, so nothing here can weigh inflation, the exchange rate, tax, labour law, platform work or any policy, and this edition names no cause for any figure in it. A reader who supplies one is supplying it from outside these data.

They do not follow anybody, and they are not a count of hiring. The two quarters are two photographs of different people three years apart, so nothing here says that an employee became an own-account worker, that anybody lost a job, or that any post was created or filled. Nor is an employer a business: the survey counts people by the status of their own job, so the rise in employers beside the fall in employees says nothing about how many people any business employs, and this edition holds no count of enterprises, payrolls or vacancies.

They do not say that unemployment fell, and it did not. The unemployment rate in these urban areas went from 6.2% of the labour force in the second quarter of 2023 to 7.9% in the second quarter of 2026, and for people aged 15 to 24 from 16.4% to 18.7%. That rate is a share of the labour force, not of the people in work, so it is never subtracted from the shares above; it is printed only to refuse the reading in which working for oneself took the place of being unemployed.

And they are not figures for Argentina. The data labels its coverage "Main cities or metropolitan areas", so every count here belongs to the urban agglomerations the survey visits. Nothing here is broken out by province, region or city, nothing is compared with another country, no reading before the first quarter of 2016 is used, and nothing in it is a forecast. These are the ILO’s processing of the survey microdata rather than INDEC’s published tables, and a reader comparing them with an INDEC release may find other numbers.

What this means for you

If you are hiring in Argentina

Seven in ten of the people in work in the urban areas this survey covers are employees, 9,554,718 of 13,541,556, so the arrangement you are offering is much the most common one, and the pool holding it is 169,851 smaller than it was three years ago. A quarter of the people in work, 3,447,070, work for themselves with nobody employed, and that is the group that has grown: by 522,761 in three years, and by 289,719 in the last year alone. Somebody coming to you from it is not changing employer, because they do not have one. The people who already employ at least one other person number 470,627, one in twenty-nine of everybody in work. These are counts of people and the status of their own job, and this read carries nothing about pay, hours, registration or the terms of any job, and nothing about any province or city on its own.

If you are looking for work in Argentina

A posted vacancy is an employee job, and the number of people holding one in these urban areas is 9,554,718, down 169,851 in three years and down 74,248 in the last year. That is the pool a posting competes for, and it has not grown with the population in work. The unemployment rate over the same three years went from 6.2% to 7.9% of the labour force, and for people aged 15 to 24 from 16.4% to 18.7%, so this is not a market in which the fall in employee numbers has shown up as nobody looking. If you are weighing work on your own account, 3,447,070 people were doing it in the second quarter of 2026, the highest of the forty-one quarters read. The survey records what status a person’s job has, never how they came to hold it, so nothing here says which route is easier or what either pays.

How this was counted

Method

  1. THE READ. Two indicators from ILOSTAT’s public data interface, which needs no key, on 10 November 2026 at 13:22 UTC. First, employment by sex and status in employment, in thousands, for Argentina, quarterly: eighty-eight quarters, from the third quarter of 2003 to the second quarter of 2026, which is the latest the series holds. Second, the unemployment rate by sex and age, percent, same country and frequency. Both queries asked for codes and labels together, so every field below is the data’s own wording and none of it is a paraphrase. [EMP_TEMP_SEX_STE_NB_Q; UNE_DEAP_SEX_AGE_RT_Q]
  2. WHAT THE DATA SAYS ABOUT ITSELF, AND THE COVERAGE THAT FOLLOWS. The source is labelled "LFS - Permanent Household Survey, Urban", the repository "ILO-STATISTICS - Micro data processing", the geographical coverage "Main cities or metropolitan areas", and the classification ICSE-93, whose categories the data names in full: employees, employers, own-account workers, members of producers’ cooperatives, contributing family workers, and workers not classifiable by status. Two consequences are load-bearing. These are the ILO’s own processing of the survey microdata and not INDEC’s published tables, so a reader comparing them with an INDEC release may find other numbers. And every figure in this edition describes the urban areas the survey visits, not Argentina as a whole: no count here is a national total, and the desk found no national series to read instead.
  3. WHERE THE COMPARISONS START, AND WHY. The data declares a break in series at the third quarter of 2003 and another at the first quarter of 2011, both labelled "Methodology revised", and it carries a remark of its own: that series published after January 2007 and before December 2015 should be used with caution, which it ties to Argentine decrees 181/15 and 55/16. Every comparison in this edition is therefore confined to readings from the first quarter of 2016 onward, forty-one quarters, all of them after both breaks and after the period the data itself cautions about. Quarters are compared only with the same quarter of another year, with the single exception named in the ratio table.
  4. THE CONTROLS. Six were run against the downloaded files and all six passed. The aggregate "Employees" equals ICSE-93 category 1 exactly, in all 177 comparisons of fifty-nine quarters by three sexes, which is what fixes the mapping of code to category. The aggregate "Self-employed" equals employers plus own-account workers plus contributing family workers exactly, every quarter. The four categories printed here equal the published total exactly in both compared quarters. Men plus women equals both sexes exactly in all four categories of the latest quarter; the sex split was read and reconciled and is not printed, because it carries no part of this finding. In each of the three windows the four category changes add to the change in the total exactly. And of the figures printed in this edition, none carries a reliability flag.
  5. DERIVED FIGURES. Every change, percentage, share and ratio here is this edition’s arithmetic on the published counts, labelled as such wherever it is printed, and computed from the published values so that a reader repeating the sum on the table lands on the same answer. That is how 434,363, 169,851, 522,761, 55,116, 26,337, 3.3%, 1.7%, 17.9%, 13.3%, 61.5%, the shares of 70.6%, 25.5%, 3.5% and 0.5%, the fall of 3.6 percentage points, and the ratios of 29.35, 30.07 and 36.08 were produced. The counts come from the data in thousands to three decimals, which is to the person, and they are printed as whole people; they remain weighted estimates from a sample survey, and the precision is the publication’s rather than the survey’s.
  6. WHAT WAS TESTED AND REFUSED. Three claims were checked against the forty-one quarters and dropped. That own-account work is at its highest relative to employee work: 36.08 per 100 is the second highest, behind 37.40 in the fourth quarter of 2020. That the employee share is the lowest on record: 70.56% is the second lowest, behind 70.53% in that same quarter, which is why a fourth quarter appears in a table of second quarters. And that there are fewer employees than ever: 9,554,718 is the fifteenth highest of the forty-one, and the series holds 10,023,469 in the third quarter of 2023. What did survive is that the number of own-account workers, 3,447,070, is the highest of the forty-one. The four shares printed for the latest quarter sum to 100.1% rather than 100.0% because each is rounded on its own, and that is printed here rather than quietly adjusted.
  7. LIMITS. Nothing here names a cause, follows a person, counts a business or a vacancy, measures pay, hours, earnings, informality, registration, contracts, tax or social security, breaks any figure out by province or city, compares Argentina with another country, or reaches behind the first quarter of 2016. The unemployment rate appears only to refuse an inference, and no figure in this edition comes from our own platform.
What the data establishes, and what it does not

What the data establishes

  • In the second quarter of 2026, 13,541,556 people were in work in the urban areas Argentina’s permanent household survey covers, against 13,107,193 in the second quarter of 2023: 434,363 more, a rise of 3.3%.
  • Over the same three years the number working for an employer fell from 9,724,569 to 9,554,718, a loss of 169,851 or 1.7%, and employees are the only one of the four categories whose count fell.
  • Own-account workers, who work for themselves and employ nobody, rose by 522,761 to 3,447,070, which is the highest of the forty-one quarters from the first quarter of 2016; employers rose by 55,116 and contributing family workers by 26,337, and the four changes add to the change in the total exactly.
  • Employees fell and the total rose in all three windows ending in the second quarter of 2026, over one year, two years and three years alike.
  • Employees were 74.2% of everybody in work in the second quarter of 2023 and are 70.6% now, and there are 36.08 own-account workers per 100 employees against 30.07 three years earlier.

What it does not

  • That own-account work is at its highest relative to employee work, or that the employee share is the lowest on record: 36.08 per 100 is the second highest ratio of the forty-one quarters read and 70.56% the second lowest share, both behind the fourth quarter of 2020, which holds 37.40 and 70.53%.
  • That there are fewer employees than ever: 9,554,718 is the fifteenth highest of those forty-one quarters, and the most the series holds is 10,023,469 in the third quarter of 2023.
  • Why anybody is counted where they are: the survey records the status of a job and never the reason, and this edition names no cause, inflation, the exchange rate, tax, labour law, platform work and policy included.
  • That the same people moved from working for an employer to working for themselves, that anybody lost a job, or that any post was created or filled: nobody is followed and a net change is not hiring.
  • That any business shrank, closed or employs fewer people: an employer here is a person, not a firm, and no headcount per employer was read.
  • That unemployment fell or that working for oneself took the place of being unemployed: the unemployment rate in these urban areas rose from 6.2% to 7.9% of the labour force over the same three years.
  • Anything about pay, earnings, hours, informality, registration, contracts, pensions, tax or social security, and anything about Argentina as a whole, any province, region or city, any other country, any reading before the first quarter of 2016, or any forecast.
Primary sources

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