One count of people, cut by how long they have been looking
In 2025 Bosnia and Herzegovina counted 173,500 people as unemployed, and Eurostat cuts that one count into eight bands by how long each person has been looking for work. The shortest is under a month and the longest is 48 months or more, with no upper limit on it. The eight add to 173,500 exactly.
Every one of those durations is a search that had not finished. The survey asks the people who are unemployed in the reference week how long they have been seeking work, so the figure is measured from the start of the spell to the day of the interview. The table records no ending, which is why nothing here can say how long a search lasts.
And to be in this table at all is to have been looking. A person without work who is not actively seeking it is outside the labour force and appears in none of these bands. So the long durations below belong to people who were still searching when they were counted.
- THE DURATION IS THE SEARCH SO FAR, NOT THE SEARCH THAT FINISHED. A labour force survey asks the people who are unemployed in the reference week how long they have been seeking work. Every band here is therefore a spell IN PROGRESS, measured from its start to the day of the interview. Nothing in the table records when a spell ends or whether it ends in a job, so no figure here says how long it takes to find work, how long anybody will go on looking, or what became of anybody counted in an earlier year.
- TO BE COUNTED UNEMPLOYED IS TO BE LOOKING AND AVAILABLE. The survey counts a person as unemployed when they are without work, available to start, and actively seeking it. Someone who has stopped looking is outside the labour force and is not in this table at all, in any band. So a long duration here belongs to a person who was still searching at the time of the survey, and the table cannot be read as a count of people who have given up.
- A SHARE OF THE UNEMPLOYED IS NOT A RATE OF THE LABOUR FORCE. 74.2% of unemployed people had been looking 12 months or more; the same people are 9.1% of the labour force, which is everybody employed plus everybody unemployed. The first divides by the unemployed and the second by a number more than eight times larger, and the two move independently: between 2021 and 2025 the rate fell from 13.4% to 9.1% while the share fell from 76.9% to 74.2% and then rose in the last two years. Eurostat publishes both units, which is why this edition can print them side by side rather than choose.
- THE LAST BAND IS OPEN ENDED, AND THE BANDS ARE NOT EQUAL. Eurostat cuts the count at under 1 month, 1 to 2, 3 to 5, 6 to 11, 12 to 17, 18 to 23, 24 to 47 months, and then 48 months or more with no upper limit. The bands cover different lengths of time, so the one with the most people in it is not necessarily the most common length of search, and this edition compares a band only with the same band in another year.
- THE SERIES IS FIVE YEARS LONG AND IT BEGINS WITH A BREAK. Eurostat flags every band of 2021 with "b", a break in the time series, and publishes no year before 2021 for this country in this table. The five years are therefore comparable with each other and with nothing earlier, and no figure in this edition is called a record, a high or a low except of those five.
- THESE ARE ANNUAL ESTIMATES FROM A SAMPLE. Each year is an estimate drawn from a survey and published to the nearest hundred people, not a register of jobseekers and not a count on a particular day. Five annual readings are five samples, and nobody is followed between them, so a share that is higher in a later year is a difference between samples and not a record of anybody staying unemployed.
So this table says how Bosnia and Herzegovina’s unemployed people were spread across eight lengths of search in 2025, how that spread reads against four earlier years, and nothing about why, about what any of these people did before, or about what happens to any search after the day it was counted.
One count of people, cut by the length of a search that has not finished, and two different things it can be divided by. A reading that takes the duration for a completed spell, or that gives a share of the unemployed and a rate of the labour force the same name, says what this table does not.
83,100 of them have been looking four years or more
83,100 of the 173,500, which is 47.9%, had been looking for 48 months or more. That one band holds nearly half of everybody counted as unemployed in the country, and it is more than the four shortest bands put together: under a month, 1 to 2, 3 to 5 and 6 to 11 months come to 44,800 people between them.
Reading the bands cumulatively, 128,700 people had been looking 12 months or more and 101,500 for 24 months or more, which is 74.2% and 58.5% of the count. Those two are published percentages and this edition prints them as published; the 47.9% is its own division of the published band by the published total.
The bands are not equal lengths, so the biggest is not the commonest. Four of them cover a few months each and the last one covers everything from four years onwards. What the last band says is that the count does not thin out at its far end, which is the opposite of what the first four bands do.
| How long they have been looking | Unemployed people | Percent of all unemployed |
|---|---|---|
| Less than 1 month | 3,800 | 2.2% |
| From 1 to 2 months | 9,500 | 5.5% |
| From 3 to 5 months | 12,400 | 7.1% |
| From 6 to 11 months | 19,100 | 11.0% |
| From 12 to 17 months | 16,800 | 9.7% |
| From 18 to 23 months | 10,400 | 6.0% |
| From 24 to 47 months | 18,400 | 10.6% |
| 48 months or more | 83,100 | 47.9% |
The same people, two denominators
The people who have been looking 12 months or more are 74.2% of the unemployed and 9.1% of the labour force. Both figures are published by Eurostat, both describe the same people in the same year, and the second is the first multiplied by the unemployment rate. That multiplication reproduces the published figure in all ten cells of the five years, which is how the two units are known to be two views of one count.
The two move independently, and over these five years they moved in opposite directions. As a rate of the labour force, long-term unemployment fell from 13.4% to 9.1% and the 24 month measure from 10.0% to 7.1%. As a share of the unemployed, the 24 month measure rose from 57.5% to 58.5%. A sentence that said long-term unemployment rose, without saying of what, would be contradicted by the same table that supports it.
The third row of the table below is derived and is labelled derived. Eurostat publishes cumulative measures at 12 and 24 months and none at 48, so the 47.9% and the 5.8% beside it are this edition’s own arithmetic on published counts, computed the same way the published rows are and marked in the row itself.
| People who have been looking | Percent of the unemployed | Percent of the labour force |
|---|---|---|
| 12 months or more | 74.2% | 9.1% |
| 24 months or more | 58.5% | 7.1% |
| 48 months or more (derived) | 47.9% | 5.8% |
Against its neighbours and the EU
Bosnia and Herzegovina’s 47.9% is the highest of the four areas read: North Macedonia reads 36.9%, Serbia 19.1% and the EU27 8.3%. The country’s share is 5.7 times the EU27 figure. This is a comparison of shares of the unemployed and of nothing else.
It is not a comparison of how many people are out of work, and Serbia is the reason to say so. Serbia counts 272,400 unemployed people against Bosnia and Herzegovina’s 173,500, more people in total and a much smaller share of them at the long end. A share of the unemployed says nothing about the size of the pool it divides.
One caution sits inside the table rather than under it. North Macedonia and the EU27 publish a "no response" band, 5,000 and 122,700 people, which is counted in the total each share is divided by; Bosnia and Herzegovina and Serbia publish none. The denominators are therefore not quite built the same way, and the table says so rather than quietly dividing by two different things.
| Reporting area | Unemployment rate | 12 months or more | 24 months or more | 48 months or more |
|---|---|---|---|---|
| Bosnia and Herzegovina | 12.2% | 74.2% | 58.5% | 47.9% |
| North Macedonia | 11.5% | 67.6% | 51.7% | 36.9% |
| Serbia | 8.7% | 39.7% | 25.9% | 19.1% |
| European Union, 27 countries | 6.0% | 31.5% | 16.4% | 8.3% |
What these figures do not say
They do not say that anybody will be unemployed for four years. The survey records how long a search has been running at the moment it is counted, never how long it lasts or how it ends, so a band of people at 48 months or more is a statement about searches in progress and not a forecast about any of them. Nor do they say that these are the same people from one year to the next: five annual readings are five samples, nobody is followed between them, and this edition names no individual trajectory.
They do not say that long-term unemployment rose. As a rate of the labour force it fell over these five years, from 13.4% to 9.1% at 12 months or more and from 10.0% to 7.1% at 24 months or more. What rose is the share of the unemployed at the long end, and the two are different measures with different denominators. Nor is 47.9% called a record: Eurostat publishes five years for this country and this table, flags the first of them as a break in the series and holds nothing before it, so 47.9% is the highest of five years and that is all this edition claims. The 12 months or more share is not at a high at all, 2021 reading 76.9% against 74.2%.
They do not say why any figure is what it is. The table records the length of a search and the count of the people, and never a reason, so this edition names no cause at all: not emigration, not the structure of the economy, not education, not benefits, not hiring practice, not the public sector and not the years before the series begins. And they do not say that the people in the long bands have stopped looking: a person who has stopped is outside the labour force and is in none of these bands, so the table cannot be read as a count of discouragement.
They do not say anything the edition did not read. Nothing here is broken out by sex, age, region, education, occupation or previous job, none of which was read for this edition; the employment rate quoted is for ages 15 to 64 while every other figure is for 15 to 74, and the two are not mixed in any sentence; no quarter is read, only annual figures; no area outside the four in the comparison table appears; nothing is a forecast; and no figure in this edition comes from our own platform.



