CENTRAL BANK OF CYPRUS
The Rise and Fall of NPLs in Cyprus: Solved, Transferred or Transformed?
By Xenios Socratous
Few banking systems in Europe, if any, have travelled such a distance in such a short time. In Cyprus, non performing loans (NPLs) were not a footnote to the great financial crisis. They rather became the crisis after the crisis.
In the mid 2010s, NPLs represented an extraordinary burden by any European standard. The ratio of NPLs to total loans rose to levels approaching, and in some portfolios exceeding, almost half of the entire loan book. Today, the picture is profoundly different. Cyprus has moved from being one of Europe’s most conspicuous NPL outliers to a banking system whose asset quality indicators align with the European Union (EU) average. That transformation is the remarkable part of the story. But it is not the whole story.
There are, in fact, two interrelated dimensions. The first is the banking sector dimension. A dramatic clean up of bank balance sheets, stronger capital, higher provisions, improved profitability and a turn to more disciplined lending. The second is the economy wide legacy debt dimension. Distressed debt did not simply disappear in its entirety. A significant portion moved outside the banks, namely to credit acquiring companies (CACs) and loan servicers, where it continues to affect borrowers, the economy and the social coherence of debt resolution.
That is why the central question is not whether the Cyprus banking sector improved. It plainly did. The better question is subtler, and more interesting: was the NPL problem solved, transferred, transformed or, in truth, all three?
Figure 1[1]
The fault lines beneath the boom: Why Cyprus Was Vulnerable
The roots of the NPL crisis were laid well before 2013. The crisis made them visible, but the underlying vulnerabilities had been building for years. Cyprus experienced a powerful credit and property boom in the years after EU accession and euro adoption. Deposits surged from the middle of the 2000s, credit growth followed, and lending to real estate and housing accelerated sharply, reaching extraordinarily high levels before the global financial crisis reached its peak. This expansion left the banking system heavily exposed to the real estate development sector, where large projects, rising land values and expectations of continued demand became closely tied to the availability of bank credit. As construction activity expanded and property prices climbed, banks’ balance sheets became increasingly dependent on the performance of developers, households and businesses linked to the property market. The result was a financial model deeply intertwined with construction, property prices, bank lending and domestic demand.
At the same time, the banking system was unusually large relative to the economy and heavily exposed to Greece, so when the Greek sovereign crisis intensified, Cyprus was hit not only through confidence and funding channels, but through direct balance sheet losses as well.
Another factor that amplified the crisis was the structural reliance of the Cyprus credit model on real estate collateral. Lending practices placed substantial weight on the appraised value of the property securing the loan overshadowing a more disciplined assessment of the borrower’s underlying repayment capacity. In the buoyant years, this appeared to offer comfort. In reality, it created a dangerous illusion of protection.
Weak provisioning methodologies reinforced the problem. Loans that were fully or substantially collateralised were often not treated as impaired, even when the borrower’s capacity to repay had weakened. As a result, losses were not recognised in a timely manner and significant under provisioning prevailed. The system remained collateral rich, but cash flow fragile.
This fragility was further masked by the widespread use of ‘extend and pretend’ forbearance measures. Loans were restructured, maturities were extended and additional collateral was often taken, despite clear evidence that many borrowers were no longer able to service their debt on a sustainable basis. These practices gave the appearance of stability, but they did not restore repayment capacity.
When property values, income and confidence fell together, collateral did not prevent losses. The protection it seemed to offer proved far weaker than the risks it had helped to conceal.
Judicial inefficiencies then amplified these vulnerabilities and made them harder to resolve. The problem was compounded by a legal environment in which foreclosure, insolvency and broader debt enforcement procedures were exceptionally slow. Very lengthy resolution timelines meant that lenders had limited incentive to pursue legal action, as enforcement could take years and outcomes were uncertain.
Instead, banks often preferred bilateral negotiations with borrowers, which frequently led to restructurings, maturity extensions or the pledging of additional collateral. At the same time, the absence of immediate consequences for non payment weakened repayment discipline. For some borrowers, delay became a strategy rather than a temporary response to hardship. This encouraged a culture in which arrears could persist, problems were rolled forward, and genuine resolution was repeatedly deferred.
A banking system can usually absorb shocks when they come from one direction like a sectoral downturn, a group of troubled borrowers, or a temporary fall in collateral values. It becomes far more fragile when several pressures arrive together.
2012-2014: The Sharp Rise and the Shock of Recognition
This is exactly what happened in Cyprus. The NPL surge of 2012-2014 was not the result of a single shock, but of several pressures arriving at once.
The Cyprus economy was already slowing when exposure to Greece and the euro area crisis began to bear down on banks and borrowers alike. Then came the 2013 banking crisis, the bail-in, restrictive measures on transactions across the banking sector and a profound loss of confidence. Household wealth was destroyed, incomes fell, unemployment rose and businesses faced a sudden deterioration in demand and liquidity.
In these difficult conditions, many borrowers who had already been under financial strain found themselves clearly unable to keep up with their debts. Others became reluctant to engage with their banks, especially in an environment that many experienced as chaotic, uncertain and hard to navigate. The disruption caused by the banking crisis, the bail-in and the measures that accompanied them was compounded by a deep sense of grievance among borrowers whose deposits, savings or other financial assets had been affected. For businesses, their owners and private individuals alike, the expectation that they should continue servicing their debts to the banking system after having absorbed significant losses on their assets was difficult to accept. This perception of unfairness fuelled anger and emerged as an important additional driver of borrower non-cooperation.
Banks, for their part, were not ready to deal with arrears on such a large scale. The tools needed to manage the problem such as restructuring capacity, reliable collateral and loan data, procedures and systems for managing different types of borrowers had to be built while the crisis was already unfolding.
But the sharp rise in NPLs was not solely a story of new defaults. It was also a story of recognition. The crisis did not simply create bad loans. It brought them into view.
On one level, the increase in NPLs reflected a real weakening in the ability of households and businesses to repay. On another level, it also reflected a more rigorous and transparent way of identifying distress. In 2014, the EBA introduced a harmonised framework for supervisory reporting, with common definitions for NPLs and forborne exposures across the EU. This was not merely a reporting change. It turned credit risk identification and classification into a more effective risk management tool, shifting the focus from the existence of collateral to the actual credit risk of the borrower. A loan could be classified as non performing not only when it was more than ninety days past due, but also when the borrower was considered unlikely to repay in full without the bank having to leverage on collateral.
The underlying principle was plain and simple. Financial distress should be recognised according to the borrower’s real repayment capacity, not postponed because of collateral availability.
Cyprus was far from alone in this. Across Europe, inconsistent classification practices had produced a patchwork of reported asset quality, one in which similar loans could be treated very differently depending on where the bank was headquartered. That lack of comparability had become a systemic problem. Andrea Enria, who served as Chair of the EBA, the SSM’s Supervisory Board and was one of the architects of European banking supervision’s post crisis reform agenda, captured the issue with characteristic directness:[2] the absence of common standards had weakened confidence in European banks as a whole, because investors could not easily tell which institutions had serious asset quality problems and which did not. In such circumstances, markets tend to punish the entire sector, the sound and the troubled alike.
Non harmonised classification was therefore a European problem, but for Cyprus it was far more profound, because the country’s outlier position made weaknesses in asset quality recognition a system wide vulnerability.
Part of the rise in problem loans reflected a genuine deterioration in borrowers’ finances. But part of it reflected something else, something necessary and, in the end, constructive. Under the new framework, loans were finally being classified with consistency, transparency and accuracy. The numbers were not getting worse so much as they were getting honest. And that mattered, because only a problem that is properly recognised can be properly addressed. What emerged was not a crisis manufactured by new rules, but the true scale of an old one, brought into the light at last.
The result was stark. Around half of the banking sector’s loan book was classified as non performing. An extraordinary level by any European standard. Crucially, however, this was not a problem confined to the banking sector. It extended well beyond the shallow waters of the financial system and into the deeper currents of the real economy.
Figure 2
Why the NPL Problem Mattered Beyond Banks
The NPL problem was never solely a banking metric. It was a macroeconomic constraint.
Large stocks of NPLs tie up bank capital, absorb management attention and impair the supply of new credit to the economy. Even viable borrowers suffered. Banks facing massive legacy exposures became more cautious, more inward looking, less willing and less capable to extend credit precisely when the economy needed a functioning credit channel.
NPLs were therefore both a symptom and a brake. They reflected the collapse of the old credit model and then obstructed the emergence of a new one.
They also created pervasive uncertainty. A household or firm carrying unresolved debt tends to postpone investment and consumption making economic normalisation slower than it otherwise would be. A highly indebted private sector may remain subdued even after GDP growth resumes.
This is why NPLs as a share of GDP are a valuable complement to the standard ratio. The NPL to loans ratio tells us about bank asset quality. The NPL to GDP ratio tells us something different but equally, if not more, important. The burden of unresolved debt relative to the economy’s capacity to absorb, restructure and resolve it. At Cyprus’ peak, both ratios were exceptional and both need to be tracked through the recovery.
Figure 3
2015-2018: The First Turning Point – Restructuring, Enforcement and Debt for Asset Swaps
The first phase of the clean up was built around operational capacity. Long before any foreclosure or insolvency reform began to have practical effect, banks had already started building the internal infrastructure needed to manage the problem themselves. The stock of NPLs was too large, too complex and too borrower specific to be absorbed through ordinary banking routines. So banks built dedicated arrears management and debt collection units, improved the quality of borrower information, applied more disciplined collateral valuations and strengthened internal governance around recovery decisions. Resolving NPLs, turned out, required an operational machinery inside the bank itself, not just a legal right to enforce.
Over this period, that machinery genuinely improved. Banks segmented borrowers more systematically, assessed viability with greater discipline and stepped up restructuring activity. Some of those restructurings held up well, particularly where the borrower’s difficulties were temporary and long term repayment capacity remained intact. Others proved less durable, especially where the underlying business model or household income had shifted for good.
Debt for asset swaps were also part of this early toolkit. Where borrowers lacked cash flow but held collateral, banks accepted real estate in exchange for debt reduction or settlement. This helped reduce impaired exposures, but created a new management challenge of its own. Banks themselves became significant holders of real estate, requiring dedicated property management units to protect and recover value.
It was onto this already functioning machinery that legal reform arrived. A modernised foreclosure and insolvency framework began to change the negotiating table. Before the reforms, the credible alternative to restructuring was often unclear for both parties. Banks could not always rely on timely enforcement, while borrowers often operated in an environment where delay could postpone difficult decisions. After the reforms, enforcement became more credible, even if it remained politically sensitive, legally contested and operationally demanding.
This was a tangible step in the right direction, but not a solution in itself. Legal reform was necessary, but it was not sufficient on its own. A foreclosure law does not automatically resolve a debt overhang, just as a new road does not guarantee that traffic will move smoothly. What made the difference was that the framework landed on top of arrears management and restructuring capacity that banks had already built to deal with the scale and complexity of the problem. The law gave that existing machinery real teeth, changing the incentives around negotiation and enforcement, and allowing banks to reap the benefits of capacity they had spent years assembling.
The key analytical challenge, once both pieces were in place, was to distinguish between three very different situations. Viable borrowers needed time and support. Non viable borrowers required orderly resolution. Strategic defaulters were exploiting weak or slow enforcement. Treating all three in the same way would have been economically wasteful and socially inequitable. Effective NPL resolution therefore required more than reducing arrears. It required judgment, consistency and the ability to separate temporary distress from permanent impairment, something that only became fully possible once credible enforcement gave banks real leverage in these conversations.
The broader European backdrop reinforced this shift. From the second half of 2010s and onwards, the EU’s action plan placed NPL reduction in a wider ecosystem of measures: supervisory expectations on provisioning, better credit underwriting standards, stronger management guidelines for non performing exposures, improved data infrastructure requirements, development of secondary markets, a blueprint for asset management companies, and benchmarking of national enforcement and insolvency frameworks. Cyprus’ domestic reforms should therefore be viewed not as an isolated national story, but as a local response increasingly nested inside a pan European strategy.
The lesson from this period is clear and sound. Restructuring and enforcement are not opposites. Credible enforcement creates the conditions for serious negotiation. But foreclosure alone is not a resolution strategy. The purpose of the framework should be to support fair, efficient and economically rational outcomes.
2018-2026: The Great Unclogging – How Cyprus Banks Cut Their NPL Burden
By 2018, Cyprus had been fighting its bad debt problem for years. Patiently, methodically, one file at a time. Banks restructured loans, chased repayments, wrote off what was gone beyond recovery, sold collateral, swapped debt for assets, and kept negotiating. It was unglamorous, grinding work. But it mattered. It forced banks to build real recovery operations. It taught them how to deal with borrowers who could no longer pay. It trimmed part of the problem. It just was not enough.
The stock of NPLs was still enormous. Too heavy, too deeply embedded, too slow to shrink through individual case management alone. What Cyprus needed was not more bedside care. It needed surgery.
From 2018 onwards, the clean up changed gear. Banks stopped treating every bad loan as a relationship to be nursed back to health or an amount to be recovered and started treating them as portfolios. Pools of risk that could be packaged, priced and moved. The stark moment was Bank of Cyprus’ Project Helix[3]. A landmark portfolio sale that marked the beginning of a new chapter in the management of NPLs. Bad loans were not just to be worked out. They could be cut out.
Other transactions followed by a number of banks. The pattern was consistent. Banks were no longer trying to climb the mountain. They were dynamiting large pieces of it.
The cooperative banking sector took a different, albeit consistent route. Less of a sale, more a controlled dismantling. Healthy banking activities remained into the mainstream system via a sale to Hellenic Bank and a large stock of troubled loans was left outside it.[4] The effect on bank NPLs was sharp and immediate. But it also revealed one of the defining truths of the Cyprus story. Bad debt does not disappear in its entirety. Part of it, moves.
The fall in NPLs cannot therefore be explained by any single cause. It was not just portfolio sales. It was not just a recovering economy. It was a set of engines and strings, running in parallel, pulling in the same direction.
Sound operational profitability returned first and quietly changed everything. When banks started generating real earnings again, they gained the capacity to absorb losses. They used profits to build provisions, accelerate recognition and write down exposures to values more closely aligned with market reality. Portfolio sales require a meeting point between a bank’s book value and what an investor will pay. As provisioning levels strengthened and became more reflective of underlying credit risk, that gap narrowed. Once it did, transactions became viable and the pace of balance sheet clean up accelerated.
Supervision also change the economics of inaction. Through measures such as the ECB’s guidance[5], the 2018 Addendum on non performing exposures[6], the CRR backstops[7] regulators reshaped the incentives facing banks. Time could no longer be relied upon as a recovery strategy. Losses had to be recognized earlier. Problem loans had to be managed, not shelved. Wait and see, on its own, was no longer an acceptable plan.
Capital and provisioning discipline quietly rebuilt the system’s shock absorbers. Higher capital buffers, IFRS 9 accounting standards and more conservative credit risk practices made Banks more capable of dealing with the NPLs. That loss absorption capacity was one of the hidden foundations of the whole clean up. You cannot solve the problem while standing on quicksand. You need solid ground first.
Table 1
Figure 4
Better lending meant the pipeline filling in behind the clean up was cleaner too. The crisis had exposed the dangers of lending against collateral rather than underlying capacity to repay. Post crisis, banks demanded more, like repayment capacity, debt service ratios, loan to value limits and cash flow analysis. New credit returned but it came back wearing a seatbelt, with default rates for new lending being contained at low levels.
Figure 5
Then came COVID-19. The first major test of the rebuilt architecture arrived in 2020 with all the subtlety of a hammer. Initially, the shock was severe and Cyprus’ response was, by European standards, efficient. At the peak of the crisis, roughly half of the entire performing loan portfolio had been placed under payment moratoria, a share that stood out as one of the highest in the EU. For a banking system that had only recently clawed its way back from the edge, that was a remarkable stress test in itself.
Yet the feared new wave of bad loans never materialised. That was not coincidental, and it was not just the moratoria buying time. It reflected something more fundamental, more profound. Banks had spent the preceding years building real shock absorbing capacity with stronger capital, higher provisions, cleaner balance sheets and more disciplined underwriting practices. When the blow came, the architecture held. A system that had once buckled under the weight of its own past absorbed one of the sharpest peacetime shocks in modern economic history without breaking. The contrast with the post 2013 years could not have been starker. Cyprus had rebuilt its banks not a moment too soon.
By the mid 2020s, the numbers told a story that would have seemed like fantasy a decade earlier. NPL ratios had fallen to levels nobody dared to predict in the aftermath of the crisis. Provisioning coverage was far stronger. Capital ratios sat among the highest in the EU. Liquidity was solid. Profitability had returned. Banks were lending again.
But this is not a fairy tale. Bad debt did not simply vanish in its entirety. Whilst some of it was resolved through restructurings, repayments, debt for asset swaps or write offs, a substantial part of what disappeared from bank balance sheets did something else. It relocated.
The Unresolved Chapter: Credit Acquiring Companies, Legacy Borrowers and the “Shadow” NPL Stock
Cyprus largely solved its banking NPL crisis before it fully resolved its private debt legacy.
CACs now hold the lion’s share of those legacy loans, with the vast majority still sitting in non performing territory. These exposures have left bank balance sheets, but they have not left the economy. Behind the cleaner ratios are borrowers still navigating unresolved obligations, active restructuring negotiations, enforcement pressure and property related outcomes. In some cases, the underlying financial difficulty has been genuinely addressed. In others, it has simply changed hands, migrating from one creditor to another, with the economic stress largely intact and the human reality unchanged.
The headline numbers, in other words, tell only part of the story.
Banks are healthier. That is a real and significant achievement, and it should not be understated. But some households and businesses remain financially distressed. The property market also remains connected to some extent with the NPL legacy through collateral sales, repossessions and the real estate portfolios accumulated by CACs and other vehicles, even as broader market conditions have strengthened in recent years. And the political debate around enforcement remains charged because foreclosure is not an abstraction. It touches homes, businesses, accumulated family wealth and deeply held instincts about what is fair.
Yet an efficient enforcement framework is not optional. It is indispensable. It is essential for the proper functioning of the credit system. Weaken it, and the costs do not disappear. They just shift. If recovery values become uncertain, if delays become chronic, or if strategic default or a weaker repayment culture is effectively rewarded, future borrowers pay the price through higher risk premia and tighter credit conditions. The line between protecting genuinely distressed borrowers and maintaining the credibility of the credit system is one of the hardest judgements of the post crisis era. Walk it badly in either direction and the consequences are real.
This is why the next chapter of the Cyprus NPL story is more complex than the first. The banking indicators reflect a remarkable turnaround. But a complete picture of the legacy debt landscape requires looking beyond balance sheets. It requires asking how much distressed private debt remains outstanding, how it is being resolved, at what pace, and with what consequences for borrowers, property values and the broader culture of credit.
The banks seem to have closed their chapter on the crisis. For many borrowers however, that chapter is still being written.
Contractual balance versus accounting value. Why the CAC stock needs careful reading
When looking at NPLs, one distinction matters more than any other. The contractual balance is not the same thing as the accounting value. This applies to banks as well as to CACs, but it is far more pronounced for CACs, where legacy NPLs make up the overwhelming share of the portfolio rather than a small pocket of it.
The contractual balance shows what borrowers owe under their original loan agreements, including accrued interest. It is the headline figure for the unresolved debt from the borrower’s side of the ledger. The accounting value tells a different story. It reflects what the loan is actually worth in the books of the holder, once expected recoveries, updated cash flow assumptions and impairments have been factored in.
That is why the two figures can diverge so sharply, and why the gap tends to be widest at CACs. A large contractual stock does not mean the holder expects to collect the full amount. Nor does a much lower accounting value mean the borrower’s obligation has quietly vanished. Both numbers earn their place in the analysis. The contractual balance speaks to the scale of the legacy debt burden still sitting in the system. The accounting value speaks to what can realistically be recovered. Because CACs are built almost entirely around legacy NPL portfolios, this distinction shapes their entire balance sheet in a way it simply does not for a typical bank. Mixing the two up can make the problem look far bigger, or far smaller, than it actually is, and for CACs in particular, it can lead to a badly distorted picture of the real economic exposure involved.
Figure 6
Lessons from the Cyprus Experience
The Cyprus experience carries lessons that travel well beyond its shores.
Recognition matters. A problem cannot be resolved before it is measured accurately. The sharp rise in reported NPLs was painful, but it was also necessary. It forced the system to confront reality rather than manage appearances. Suboptimal loss reporting and improper loan classification that was the case under the old framework may defer pain, but they rarely produce solutions.
Legal framework matters; Credibility matters more. Foreclosure and insolvency reform can change the incentive structure, but only when the framework is seen as credible, predictable and capable of delivering outcomes within a reasonable timeframe. A law that exists on paper but does not influence borrower and lender behavior is not yet a resolution mechanism. It is an expression of intent.
Collateral is not a substitute for repayment capacity. A property held as security can limit losses when things go wrong. It cannot make a non viable borrower solvent. The pre crisis model systematically overestimated the protection offered by collateral and underestimated the importance of sustainable borrower cash flows. That lesson should not need to be relearned.
Supervisory pressure is a policy instrument, not just oversight. Banks acted faster and more efficient when capital requirements, provisioning expectations, accounting standards and market discipline made inaction expensive. Incentives that raise the cost of delay are as important as the technical tools available for resolution.
Portfolio sales are powerful, but partial. Selling NPLs repairs balance sheets, rebuilds investor confidence and frees up capital for new lending. But the risk does not evaporate. It relocates. The borrower still exists. The debt still exists. The stress still exists, now managed by a different hand. Measuring success only at the banking sector level risks missing half the picture. This is precisely why tracking the progress of CACs matters. The loan does not stop being real once it leaves a bank’s balance sheet, and neither does the borrower behind it. Following what happens next, how CACs manage these portfolios, how much is actually recovered and how borrowers fare under new management, is what turns a partial picture into a complete one.
Borrower discipline is a public good. When enforcement is weak and strategic default goes effectively unchallenged, the cost is not absorbed by the lender alone. It spreads. It is priced into the next loan, passed on to the next borrower and embedded in the lending standards of the next cycle. Future borrowers who had nothing to do with the original crisis inherit higher risk premia, tighter credit conditions and a weaker repayment culture. A foreclosure and enforcement framework that lacks credibility, speed or finality does not protect borrowers. It protects those who have decided that the system will not catch up with them.
NPL resolution is not a single policy. It is an ecosystem: Recognition, provisioning, restructuring, enforcement, repayment discipline, functioning secondary markets, sound governance. Remove any one of these and the others work less well. The Cyprus story was ultimately a story of multiple forces converging. Imperfectly at times, unevenly at others, and often painfully. But converging nonetheless. That convergence was not inevitable. It was the product of choices, pressures and, eventually, a shared understanding that the alternative was worse.
Conclusion: Real Progress, and Unfinished Business
Cyprus’ NPL story is a genuine success story. But not a tidy one.
The banks are stronger. Ratios that once stood at crisis era extremes have fallen to European averages. Provisions are higher, capital is more robust, and the Cyprus banking system no longer occupies the uncomfortable position of European outlier it once did. That achievement deserves to be cherished since it was not handed down. It was earned through years of painful restructuring, relentless supervisory pressure, substantial assets disposals, legal reform and considerable operational effort by the banks themselves. It was hard work, and it paid off. The banking problem has been largely resolved. The private debt problem however is still very much alive.
As such, the next chapter of this journey extends beyond the confines of the banking sector. It is about something harder and less clear. It is about resolving legacy debt fairly and efficiently, sustaining a culture of repayment that benefits everyone, protecting viable borrowers in genuine distress while not rewarding those who strategically decided to stop repayments. And above all, ensuring that as memories of the crisis soften and fade, the instincts that produced it do not quietly return.
Because the fall in NPLs is not simply a story about bad loans becoming fewer. It is a story about a country learning, at considerable cost, and not without pain, that credit is not wealth unless it can be repaid.
That lesson was expensive. It would be a shame to forget it.
The views expressed are those of the author and do not necessarily reflect those of the
Central Bank of Cyprus or any other institution or organization.
[1] As per European Banking Authority (EBA) definition, i.e., including cash balances and exposures to credit institutions and central banks
[2] Keynote speech by Andrea Enria, Chair of the Supervisory Board of the ECB, at the ECB Supervisory Reporting Conference 2023
[3] Announcement by Bank of Cyprus (28 August 2018): Agreement for sale of a portfolio of non-performing loans
[4] Announcement by Hellenic Bank (03 September 2018): Hellenic Bank completes the acquisition of certain assets and liabilities of the Cyprus Cooperative Bank
[5] Guidance to banks on non-performing loans
[6] Addendum to the ECB Guidance to banks on non-performing loans: supervisory expectations for prudential provisioning of non-performing exposures