Phone bills fell. Where did the money go, and where did the workers go?

Premise

Suppose that when a price falls, the money and the people go their separate ways.

This piece starts from that assumption and looks at what is happening now in its light. It is an assumption, not a conclusion.

Seven countries, measured the same way

A team of economists measured what it costs to lose your job in seven European countries — Denmark, Sweden, France, Austria, Italy, Spain and Portugal — using the same method in each. All seven keep registers that link workers to the firms that employ them, and the data were harmonised before any comparison was made.

In Denmark and Sweden, people lose the least income after a layoff. In Italy, Spain and Portugal, they lose three times as much.

France and Austria sit in between.

So the same event — a factory closes, a division is cut, your name is on the list — comes with one of three price tags, depending on which country you are in when it happens.

All seven countries measure these losses, so the gap does not come down to whether a country counts them.

The study points to two mechanisms. In southern Europe, people who lose a job are less likely to find another one, and that explains much of the gap. And in all seven countries, a large part of the loss comes from something more specific: the wage premium tied to the old employer. Some firms pay more than others for the same work, and you don’t take that premium with you when you leave.

What this has to do with your phone bill

Telecoms have got cheaper in Europe. Prices for information and communication are 26.6% lower than in 2010. The share of household spending that goes on telecoms fell from 3.10% in 2005 to 2.30% in 2022.

Jobs went down too. The domestic workforce of European operators shrank from 550,000 in 2019 to 493,000 in 2022. Worldwide, operators employed 4.34 million people in the fourth quarter of 2025, 1.9% fewer than a year earlier.

The two happened over the same period. This piece does not claim that falling prices caused the fall in headcount. Prices can fall for several reasons: better ways of working, thinner margins, cheaper inputs or lower taxes, or a change in what is being sold.

Something else, though, did not happen.

When a company spends less on wages, you would expect the saving to show up as profit. MTN Consulting, which tracks telecom headcount quarter by quarter against company results, finds no direct link between job cuts and wider margins — not even allowing for a time lag, and whether profit is measured as operating profit or as EBITDA.

The same analysis says where the savings went instead. They were offset by other costs, chiefly depreciation — the part of an asset’s value that is written off each year. The more network an operator builds, the bigger that charge becomes.

In short, the money moved from wages into equipment.

Then the trail goes cold

Say a household stops paying part of its bill. That money goes somewhere. But where?

Europe’s shared survey can’t tell us. The Household Budget Survey does not follow the same households over time. Each round shows how spending is split, and the rounds are not fully comparable with each other. There is no way to match a household before the saving with the same household afterwards.

The survey was not designed for this question. That is not a flaw: it was built to do a different job.

The public figures don’t show where the workers went either

Some sectors did grow over the same period. The number of ICT specialists goes up each year — by 2.6% between 2024 and 2025. Jobs in renewable energy grew by 50% between 2015 and 2023. Health, which accounts for more than a tenth of Europe’s GDP, is linked to an estimate of two million new jobs.

Did the people who left telecoms end up there?

The EU’s published table of labour market flows doesn’t say. It shows nine transitions between three states — employed, unemployed, and outside the labour force. Moving from one job to another in a different industry is not among the figures it presents.

Researchers can work it out from the microdata: the survey records the sector of a person’s previous job, and consecutive quarters can be linked. What is missing is a version that an ordinary citizen can open.

So the problem is not whether this can be measured, but what public statistics choose to show us.

Two Brussels municipalities

Belgium offers a case small enough to see up close. In December 2024, a fourth mobile operator entered the Belgian market, and the regulator then published an international price comparison. For mobile plans of up to 10 GB a month, Belgium moved out of the expensive end of the table, and rival operators changed their prices too.

Fixed internet was a different story. The newcomer’s price was the lowest compared with neighbouring countries — but it was available only to people living in Anderlecht and Sint-Jans-Molenbeek. On the fixed market, the regulator wrote, the other operators had shown no reaction yet.

The cheaper price stopped at the boundary of two municipalities.

The two markets do not cost the same to break into. For fixed lines, the biggest cost is civil works: digging up the streets. Bringing fibre to a home costs 200–250 euros per household in a dense city, and 1,000–1,500 euros or more further out. Mobile has no comparable cost.

That is how the costs stack up. This piece is not claiming that they explain where the newcomer chose to build; the two facts are simply set side by side.

Measuring a loss doesn’t shrink it

Statistics can show what people paid. What sets that price is the labour market and its institutions.

The seven-country study shows this most clearly. Each of those countries could see the loss, and in some of them it was still three times larger than in others. Measuring a loss lets you see it; it does not make it smaller.

Wider research on job loss points the same way. Of those who find new work, 56% join a firm that pays less. In some countries, earnings halve in the year of the layoff and are still about 10% lower four years later.

What sets the price is how institutions are designed — and who, in that design, ends up absorbing the shock.

What remains open

This piece does not claim that AI is the cause. Telecom headcount was falling before today’s AI systems existed. It is fair to ask whether AI speeds up something that was already under way — but this piece goes no further than asking.

Nor is this just a telecoms story. Wherever prices fall and the way people work changes, the same question comes up:

Where did the money that was saved go? And where is the person it used to pay?

There is not yet a public table that answers the second question.

Where I stand

I don’t judge a transition only by the total number of jobs. I also look at whether people can actually follow the money to where it went.

That is not a recommendation. It is simply what I am watching.

What I got wrong along the way

While researching this piece, I stretched a narrower point — that these moves don’t show up in public statistics — into a much bigger one: that they are not counted anywhere. Studies built on records that link workers to their employers proved that wrong. This piece uses only the evidence that held up after the correction.

What this piece has established

• The EU’s Household Budget Survey does not follow the same households over time, so it cannot show a household before and after a saving
• The EU’s published table of labour market flows does not show people moving between industries
• In seven countries measured with linked data, the earnings lost after a layoff differed by up to a factor of three
• The study puts forward two main mechanisms: the chance of finding a new job, and the loss of a wage premium tied to the old employer

What this piece does not claim

• That measuring the losses made them smaller
• That AI is the cause of this change
• That falling prices caused the fall in employment