How the 2012–13 Cyprus financial crisis set back wages and jobs for nearly a decade

What happened

A property boom, a pile of Greek bonds, and two banks that found themselves in the middle of it.

Cypriot banks had grown vast relative to the country: by 2011 their assets were around eight times the island's annual output, swollen by foreign deposits and offshore business. They had put that money to work in two places. In Cyprus, they lent heavily against property during a boom, and as prices later fell by roughly a third, the share of loans that stopped being repaid climbed from under 10% in 2009 to more than 45% by 2013 — and dealing with those non-performing loans would drag on for years. Abroad, they had bought Greek government bonds on a scale no other banking system in the euro came close to: about €4.7 billion of them, well over a fifth of Cyprus's entire GDP.

That second bet is what detonated first. Through 2010 Greek ten-year bonds paid something like 10 to 12%, against about 3% on the safe German equivalent — the highest yields of any government in the euro, precisely because the market already doubted Greece would pay. Investigators later found that the Bank of Cyprus more than quadrupled its Greek holdings that year, buying around €2 billion, and its own internal emails described the reasoning bluntly: with bad property loans eating into profits at home, the bank reached for the richest yield it could find abroad to make up the difference. It was reaching for Greek debt to cover the hole that Cypriot property had already opened — crazy, reckless, naive? You can decide on that.

When the eurozone restructured Greece's debt in 2011 and 2012, private bondholders took a nominal loss of around 53%, and Cyprus's banks absorbed roughly €4.1 billion in write-downs: the Bank of Cyprus around €1.8 billion, Laiki €2.3 billion. In a single stroke, losses worth more than 22% of Cyprus's annual output were carved out of two banks that did not have the capital to withstand them. Whether they should ever have been allowed to concentrate so much in one struggling neighbour became the subject of a parliamentary inquiry and years of litigation; the holdings themselves were legal, and no agreed account of who was to blame has ever emerged.

The rest followed from there. The ratings agencies, citing the banks' Greek exposure and the cost the state would face in rescuing them, cut Cyprus's credit rating to junk, which shut the government out of international markets just as it needed to borrow to prop up its lenders. A bank problem and a government problem had become the same problem, each pulling the other down.

How they got there was never fully established. The central bank had brought in a restructuring firm, Alvarez & Marsal, in the summer of 2012 to find out. When its investigators went looking, the email records of two former Bank of Cyprus executives covering 2009 to 2012 — the years of the bond buying and the property lending — had largely gone, and software for deleting email that was not part of the bank's standard equipment was found on their machines. The investigation itself then became contested: the central bank's contract with the firm carried an €11 million success fee that reportedly never went before its own board, and after documents about the arrangement leaked the attorney-general opened an investigation into the central bank governor. No charges followed. How two banks came to fail has been argued over ever since rather than answered.

What followed was the part that has stayed with most people: rather than a conventional bailout, Cyprus agreed a bail-in: Laiki was wound up, and large uninsured depositors at the Bank of Cyprus had their money converted into shares in a bank that no one had any confidence in. For the first time anywhere in the euro, limits were put on what people could do with their own money. While the banks stayed shut, cash machines at the two failing lenders were rationed to €100 a day. When the banks finally opened on 28 March, the limit was €300 a day per person per bank, cheques could not be cashed, and anyone leaving the island could carry no more than €1,000, with officers at Larnaca airport under orders to confiscate the rest. Credit stopped, confidence collapsed, and the damage became more tangible for the average Cypriot, as restrictions seeped out of the banks and into the labour market.

Add up what the rescue actually cost and the total ran to more than the island's entire annual output. The troika lent €10 billion, and that money went almost entirely to the government itself, to help it redeem its own maturing debt and cover the ordinary budget now that the credit downgrade had shut it out of borrowing normally. Cyprus itself was left to find some €13 billion more to cover the rest.

How Cyprus's own share was raised is its own tangled story: higher taxes, a slice of the central bank's gold reserves, a rollover of domestic government bonds, privatisations that mostly did not happen, and a formal default imposed on junior bondholders, on top of the depositor losses below. The total required moved several times over the three weeks the deal was negotiated, and even the officials involved later admitted the successive figures were not strictly comparable. That is worth an article of its own. This one stays with the impact on jobs, salaries, and the average Cypriot.

How far it fell, and how long it stayed down

Jobs and pay, measured against where they stood before the crash.

Start with the scale of the job losses. Employment peaked at just under 400,000 people in 2011 and bottomed at 358,000 in 2015 — a fall of 10%, or roughly 40,000 jobs, in a workforce of just over 430,000. It took until 2018 for the number of people in work to climb back to where it had been, seven years after the peak.

Pay followed a different rhythm. Wages did not collapse so much as sag, and then stay sagging, drifting down for four straight years before finally turning. Switch between the three views below and the shape of each one is worth watching.

Employment, average wage and median wage, each set to 100 at its pre-crisis peak (2011 for jobs, 2012 for pay). Employment from CyStat's Labour Force Survey; earnings from CyStat's Social Insurance records. The shaded band marks 2012–2016.

The average wage fell 5.3% from its 2012 peak and was back where it started by 2020. The median wage — what the worker in the middle actually takes home — fell 6.9%, and did not recover until 2021. The typical worker lost more than the average did, and waited an extra year to get it back.

What it cost

Counting the jobs that never existed, and what they would have paid.

The usual way to price a crisis is to imagine the good times continuing: take the wage as it stood before the crash, grow it forward at some plausible-sounding rate, and measure the gap. The trouble is that the answer then depends almost entirely on the rate you picked, and you can move it by a factor of ten without anyone noticing.

There is a stricter way. Rather than assume wages would have risen, assume nothing at all. Suppose only that the number of people in work had stayed exactly where it was in 2011. Employment peaked at 398,218 that year. It fell to 385,227 the next, then 365,083, bottoming at 358,205 in 2015. Subtract each year from the peak and you have the jobs that did not exist: about 13,000 in 2012, 33,000 in 2013, 40,000 at the worst of it. Every one of them was a wage that was never paid to anybody.

But which wage? That depends on what kind of work actually disappeared.

Employment in each sector compared with its 2011 level, year by year (CyStat Labour Force Survey). Bars to the left are jobs below the 2011 level; bars to the right, above it. Construction is highlighted.

Drag through the years and two things stand out, neither of them surprising: construction takes the biggest hit. It shed 20,779 jobs by 2014 — 45% of everyone working in it, a workforce of 46,000 cut to 25,000 — and was still nearly 15,000 down in 2017, long after the rest of the economy had begun healing. The second is that the recovery was never even. By 2016 hotels, health and professional services were all above their pre-crisis level while construction and domestic work were still deep in the hole.

This matters for how we calculate the estimated cost of the crisis. What was destroyed was mostly building sites, shop floors and household work, so weighting the missing jobs by the sectors they actually came from, year by year, gives an average of about three-quarters of the national wage — and it falls further as the recovery goes on, because the lower-paid jobs are the ones that stay missing. That is what each vanished job is valued at below, rather than pretending a lost bricklayer earned what the average Cypriot earned.

Employment against a flat 2011 baseline. The shaded area is the shortfall: jobs that did not exist. Multiplying each year's shortfall by what those particular jobs paid — sector-weighted, and recalculated each year — gives the wages never earned. Employment from CyStat's Labour Force Survey, earnings from Social Insurance records.

On that reckoning the crisis destroyed about 175,000 job-years — as if 175,000 people had each spent a full year out of work — and with them roughly €2.9 billion in wages that were never paid to anyone. That is a cumulative figure spread across six years, in an economy that produced around €18 billion a year, so it amounts to something like one euro in forty of everything the island earned over that stretch. And it is a floor, not an estimate: it assumes the economy would not have created a single extra job in six years. Nudge the slider even slightly and the figure climbs.

A number that large needs checking against independent data. There are two obvious tests.

The first is unemployment, which is measured separately. If those jobs really vanished, the people who held them should show up somewhere, and they do. Unemployment climbed from 7.9% in 2011 to a peak of 16.1% in 2014, and the rise in people out of work tracks the fall in jobs closely. In 2014 this model says 35,473 jobs were missing; the unemployment figures, compiled separately, say 35,557 more people were out of work than in 2011. Two datasets, built from different surveys, agree to within a hundred people: the missing jobs show up, almost one for one, as people out of work. One caveat belongs here: the labour force did shrink, by about 4.5% from its 2012 level, as some people gave up looking or left the island, so a slice of the loss shows up as withdrawal rather than measured unemployment.

The second test is whether the flat-jobs baseline is too generous. Measure the damage the other way round — as excess unemployment above a normal rate — and you get 152,000 job-years if you treat 2011's already-elevated 7.9% as normal, or 225,000 if you use the 5% the island ran before the crash. The figure here, 175,000, sits between the two, so again we're validating the model.

The obvious objection is that the 2011 peak was itself inflated by a property boom, so some of those jobs were never going to survive anyway. That is a fair worry, and the answer is in the later data: employment passed its 2011 level again in 2018 and has kept climbing since, reaching over half a million people in work by 2025, a quarter above the supposed bubble peak. Whatever else 2011 was, it was not a ceiling the economy could never reach again.

Who actually paid for it

Looked at through wages alone, the crisis seems milder than it should. The reason is that prices were falling at the same time.

This is the difference between the number on your payslip and what that number actually buys. Between 2013 and 2016 prices in Cyprus fell by more than 5% in total — and not by accident. A country with its own currency would have devalued it to claw back competitiveness; Cyprus, inside the euro, could not, so the bailout programme pursued the same end by the only route left — pushing wages and prices down directly, a process economists call internal devaluation. The falling cost of living was, in other words, part of the same treatment that was holding wages down. Across the whole decade from 2013 to 2022, Cypriot prices rose by just 6.3%.

All of which holds, with one large exception that never appears in wage or employment figures. The rescue was paid for partly by depositors. To refill the Bank of Cyprus with capital, savers with more than €100,000 in the bank had 47.5% of everything above that threshold converted into shares — the first €100,000 in any account was protected by EU deposit insurance and left untouched, but anything beyond it was not — and at Laiki, the second-biggest bank, the unprotected deposits were wiped out as the bank was wound down. Bank shares and bonds went with them. So a person could have kept their job and their salary all the way through the crisis and still have lost a substantial part of their life savings in a single weekend in March 2013. The burden that does not appear in any of the charts above fell on wealth rather than pay, and it fell hardest on the people who had managed to accumulate some.

For everyone below that threshold, though, the picture holds: the crash did not fall evenly and thinly across all wages. It fell very heavily on the 40,000 people pushed out of work, somewhat on those whose savings were bailed in, and comparatively lightly on the majority who kept their jobs and their deposits and watched prices fall alongside their frozen pay.

The ones who got out first

A bail-in can only take what is still in the account on the day of reckoning.

The €100,000 threshold tells us that those with savings above 100K in a single bank were liable to pay for the bail-in, but it doesn't explain who actually paid, because between the first bailout proposal and the moment the money was frozen there was a window, and some people were in a position to take advantage of that.

The banks shut on 16 March, but cash machines kept dispensing and electronic transfers continued for exceptional reasons — medicine, humanitarian supplies, jet fuel. Both banks ran branches in London where no limits applied at all, and Bank of Cyprus owned 80% of a Russian lender, Uniastrum, which was similarly unrestricted. Anyone with the standing to use those routes, and critically the knowledge about what's coming, could still move money while a regular depositor in Nicosia was rationed to €100 a day.

And money did move, and in large quantities. Savers pulled €4.1 billion out of the Bank of Cyprus over the first half of 2013, with capital controls in force for almost all of it. Der Spiegel reported that substantial sums had left the two banks even before the first bailout was signed in the small hours of 16 March, despite central bank orders freezing those accounts.

People felt wronged — as they should have felt. Lists circulated of companies and individuals who had moved money out in the final days, and parliament opened investigations into withdrawals made before and after the ban. It emerged that members of President Anastasiades's extended family had wired €21.5 million abroad days before the crisis came to a head, and that clients of the law firm he founded, and left two days before taking office, had also moved money out ahead of the bail-in. The anti-money-laundering unit investigated after the president himself asked it to, and cleared the firm in 2019 — a clearance that investigators revisiting the case have since described as flawed and incomplete. In June 2026 the Anti-Corruption Authority recommended that the Attorney General open criminal investigations into the firm, its partners and the former president. The firm and Anastasiades have denied any wrongdoing throughout, nothing has been proven, and thirteen years on that is still where it sits.

Regardless of whether there is any wrongdoing, whether it's a conspiracy, or whether some powerful individuals did manage to escape with their money, a bail-in punishes whatever is in the account on the day. So anyone who managed to move their money paid nothing at all — whether they moved on a tip, on a rumour, or on ordinary caution. The people who took the 47.5% hit were the ones with no particular reason to think they should run, no branch in London to run to, and no way of knowing the door was closing, in other words, the ordinary Cypriot.

The gender pay gap that narrowed over the crisis

One statistic improved right through the crisis: the gender pay gap. Measured on the median, it fell from 21% in 2010 to 9% by 2025. It is easy to read that as women steadily gaining ground, and in the recovery years that is exactly what happened: since 2016 the female median has risen 35% against 28% for men.

But look at when the gap closed fastest, and we see it was not because women were gaining ground, but because men were losing more. Between 2012 and 2016, at the depth of the crisis, the male median wage fell 9.7% while the female median fell just 3.5%. And by now the reason should be obvious: construction, which shed more than 20,000 jobs and 45% of its workforce, is overwhelmingly male. The sectors that were growing through the same years — health, social work, professional services — are typically not. The collapse that drove the €2.9 billion and the collapse that closed the gender gap are the same collapse.

The gap between male and female gross monthly earnings, as a percentage of male earnings, on both the mean and the median, 2010–2025 (CyStat). The shaded band marks the crisis years, when the gap narrowed fastest — largely because men's pay fell hardest. Note this is a monthly gap, so it reflects both what people are paid and how many hours they work; the headline "gender pay gap" usually quoted for Cyprus is Eurostat's hourly measure, which stood at 16.1% in 2011 and has since fallen below the EU average.

The two lines also behave very differently. On the mean, the gap has barely shifted in fifteen years, from 19% to 14%. On the median it has more than halved. The difference tells you the remaining gap is concentrated at the top of the pay scale rather than the middle: among typical earners men and women are now close to parity, while among high earners the men are still, by some distance, the ones being paid.

The long shadow

The bail-in itself was over in a fortnight. Repairing the labour market took nine years, and the repair was paid for by people who had nothing to do with the bonds or the property lending.

The bill, then: nearly three billion euros in wages never paid, four consecutive years of falling pay, and a median wage that did not see its 2012 level again until 2021 — nine years later. And a gender gap that closed partly because men's wages fell faster than women's, which can't be sold as a story of progress.

Cyprus has since recovered, and then some: more people are in work today than at any point on record. But the recovery took long enough that a person who left school in 2013 spent the whole of their twenties in its shadow.