Banks under pressure

This article is part of In conversation about sustainable finance & emission reduction systems, a new series by Diego Balverde. You're reading volume five of the The collateral crisis series. Here is volume four

Prologue

Banks entered this phase with balance sheets that still look sound, but they are no longer operating on the same terrain that delivered stability over the previous decade. Today they carry exposures to firms facing more expensive energy, weaker logistics, thinner margins, costlier debt and assets that are increasingly being revalued through climate exposure, transition pressure and collateral quality. The problem is not that the European banking system is visibly broken. The problem is that the economic base supporting the quality of its assets has become more unstable and more expensive at the same time. At the end of the third quarter of 2025, banks in the EU and EEA still reported a CET1 ratio of 16.3%, an LCR of 160.7%, an NSFR of 126.8%, total assets of 29.1 trillion euros, non-performing loans of 373 billion euros and a return on equity of 10.7%. Those are solid numbers on the surface. 

But they only remain reassuring if deterioration stays contained. That can no longer be assumed in a world of high geopolitical risk, more selective credit and productive assets under rising cost pressure.

The appearance of strength is hiding a deeper tension

At first glance, the system still looks resilient. A CET1 ratio above 16%, an NPL ratio around 1.8% and profitability above 10% suggest banks are comfortable. But those numbers should not be interpreted as proof of invulnerability. They are only a good starting point relative to a potentially more hostile environment. European banks manage more than 29 trillion euros in assets. On a balance sheet base that large, even minor percentage deterioration can turn into very large absolute losses. If a limited share of corporate portfolios begins to suffer because margins are being compressed by energy, freight, rates and weaker demand, the impact may look gradual at first and then accelerate when refinancing becomes more expensive and cash flows are no longer sufficient to cover debt service comfortably. Banks do not need another 2008-style event to face pressure. They only need a sufficiently broad deterioration in borrower quality across multiple sectors at once.

That is the dangerous blind spot. During 2023 and 2024, a large part of the improvement in bank profitability came from higher interest rates, because loan income adjusted faster than deposit costs. That temporarily boosted net interest margins. But that effect is not permanent. Once deposit competition intensifies, wholesale funding remains expensive and provisioning begins to rise, those gains can reverse. The issue is not that banks became safer because the real economy strengthened. The issue is that they made more money for a while while the deterioration of underlying assets remained delayed. If that delay ends, then what currently looks like strength may start to look like insufficient protection against a more hostile system.

The main risk is not current default; it is the direction of deterioration

Current non-performing loan ratios remain low by historical standards, but the system is already showing stress where it matters most. NPLs in the EU and EEA banking system stood at roughly 373 billion euros in mid-2025, with an overall ratio around 1.84% and coverage near 41.7%. That looks manageable until the composition is examined. Consumer credit was already showing an NPL ratio of 5.4%, SME loans 4.6%, and gross inflows into default approached 110 billion euros in the first half of 2025. In other words, this is not a system without pressure. It is a system still managing pressure. Banks are containing deterioration through recoveries, restructurings, write-offs and cures, but risk is still entering the system.

That changes how current stability should be interpreted. If a 10-billion-euro corporate portfolio exposed to energy-intensive borrowers sees average EBITDA fall 15% to 20%, debt-service coverage can deteriorate rapidly. If refinancing costs increase, insurance rises, demand weakens and the value of collateral becomes less reliable, default does not need to jump suddenly to generate a problem. It is enough for more loans to migrate into Stage 2, meaning they are not yet defaulted but already show significant deterioration in risk. And that is where the real warning sits. In the third quarter of 2025, Stage 2 loans still represented 9.3% of total loans in the EU and EEA banking system. That means nearly one in ten loans was already under elevated risk surveillance. That is not a crisis headline, but it is not normality either.

Funding costs are up while asset quality is under pressure

Banks suffer most when funding becomes more expensive at the same time that assets become weaker. That is precisely the risk of this phase. Liquidity ratios such as the LCR at 160.7% and the NSFR at 126.8% suggest the system is not short of regulatory liquidity. But averages conceal fragility. Dollar liquidity coverage has been tighter for a meaningful share of institutions, and part of the liquid asset buffer has shifted away from central bank reserves toward sovereign paper. 

None of that is alarming in isolation. 

But it shows that liquidity is no longer sitting on the same base that characterized the years of extraordinary monetary accommodation.

At the same time, banks are not funding themselves in a zero-cost environment anymore. They pay more to retain deposits, more to issue wholesale debt and more to maintain balance-sheet flexibility. If they also need to increase provisions because risk is rising, then profitability gets compressed from both sides. A small decline in net interest margin may look manageable in percentage terms, but on very large balance sheets the absolute effect is substantial. With NIM around 1.58%, cost-to-income around 52.3% and cost of risk around 0.47% in late 2025, the system still looked functional. 

But those metrics are not immune. 

If margins narrow, cost of risk rises and capital requirements become more demanding, then credit capacity begins to shrink even without an official lending freeze. That is the invisible tightening. Banks do not need to announce they are lending less. They simply lend more selectively, demand more collateral, charge more spread and shorten maturities.

When collateral changes, the bank changes

The most important mutation of this stage lies in collateral. Banks can continue lending as long as they believe that if a borrower fails, the asset backing the loan retains value and liquidity. That assumption is no longer automatic. An energy-inefficient building, an industrial facility exposed to volatile utility costs, a borrower tied to sectors with difficult transitions or a business whose margins depend heavily on expensive transport and carbon exposure is not the same kind of collateral it was five years ago. Central banks have already understood this. Climate factors are increasingly being introduced into collateral frameworks in order to reflect transition vulnerability and the possibility that the same asset may deserve a lower valuation under future stress.

This is a profound shift because it changes banking language itself. Climate deterioration used to sound like a long-term or reputational issue. Now it is entering the mechanics of discounting, haircuts, eligibility and collateral resilience. If central banks incorporate climate-related valuation factors, commercial banks cannot continue assessing certain exposures using old assumptions. And if supervisors require stronger climate-risk governance, then the entire system receives the message that these risks are no longer optional or decorative. They are prudential. That means weaker collateral is no longer only an environmental concern. It becomes a financial one.

Implementation and value capture

In this environment, the only strategic path for both banks and borrowers is not to wait for shocks to fade but to reduce operational fragility in measurable ways. This is where efficiency becomes a balance-sheet defense. If an industrial company reduces energy use by 15%, cuts idle time by 20%, improves logistics efficiency by 10% to 15% and documents lower volatility in its cost structure, it does not merely raise margins. It improves cash flow stability. 

That immediately changes how a bank sees the borrower.

Lower spread, better internal credit treatment and greater willingness to refinance can follow.

That is where models like BalGreen have real economic meaning. 

Not because they sound sustainable, but because they identify where money is leaking, where energy is being wasted, where throughput is being lost and how much cash flow can be recovered. 

A port operation that reduces idle time by 25%, energy consumption by 15% and asset rotation inefficiencies by 20% is not simply becoming more productive. In a debt structure of 100 million euros, a reduction in financing cost of 50 to 150 basis points means between 500,000 and 1.5 million euros annually. If those gains are traceable and supported by measured reductions in emissions and process volatility, then the asset becomes more defensible not only to the bank, but to the regulator and to the market. In this phase, documented efficiency is a silent form of recapitalization.

Analysis

The debate can no longer be reduced to “banks look strong” or “there is no crisis yet.” 

The real questions are harder. How much of current banking strength still depends on extraordinary rate-cycle profitability rather than on a genuinely stronger real economy?

How long can low NPL ratios remain reassuring if deterioration is already entering through consumer credit, SMEs, industrial margins and collateral quality? 

What happens if the next shock does not hit one sector but several at once through energy, freight, rates and weaker demand? How many banks are still valuing corporate exposures as if expensive energy, disrupted logistics and transition risk were marginal rather than central? 

Is it still credible to assume that low default today guarantees low loss tomorrow?

What happens when Stage 2 pressure stops being a yellow light and begins to migrate into broader default? How much of today’s lending is still secured by assets whose book value looks stable while their real collateral resilience is worsening? 

Can banks continue acting as neutral transmitters of credit if the collateral they receive is no longer neutral? 

And are we not perhaps moving toward a banking system that may not fail through classic leverage excess, but through a slower, more technical and less visible combination of weaker borrowers, softer collateral, higher funding costs and tighter supervisory expectations?

The deeper question is even more uncomfortable: is the banking system still financing the future, or is it still financing an economy designed for an energy, logistics and environmental cost structure that no longer exists? 

Because if it continues to finance the past using the assumptions of the past, the pressure will not disappear. It will accumulate.

Conclusion

My conclusion is clear. Banks are not yet the visible epicenter of the crisis, but they can become the main transmission channel of systemic deterioration if they continue to finance with old logic an economy whose costs, risks and balance-sheet structure have already changed. The problem is not only the default. The problem is the future quality of the asset, the durability of the cash flow supporting it and the capacity of the collateral to preserve value in a world that is more expensive, more regulated and more unstable. In that new stage, credit will not disappear overnight. It will become progressively more selective, more expensive and more demanding.

That means the line between staying inside the system and beginning to fall outside it will not be determined by narrative. It will be determined by the ability to demonstrate efficiency, operational resilience and measurable risk reduction. Whoever can turn friction into savings, savings into cash flow and cash flow into stronger financeability will remain within the credit system. Whoever cannot will discover too late that the bank did not stop lending out of ideology, but out of balance-sheet necessity. And when the balance sheet decides, the discussion ends.