Interview with Bloomberg

Interview with Pedro Machado, Member of the Supervisory Board of the ECB, conducted by Nicholas Comfort and Esteban Duarte on 30 October 2026

30 October 2026

What are the benefits and risks that you see from securitisation for banks?

We welcome the proposal for reviving the securitisation market as presented by the European Commission. Securitisation, if enabling properly structured significant risk transfers, is a legitimate and efficient instrument for banks to manage their capital and credit risk. At the same time, certain safeguards have to be observed to boost lending with the capital relief you generate. You could then foster lending to address the financing needs that Europe is confronted with. We also see it as a way to create a certain dynamic in capital markets in Europe. The issue with securitisation is, of course, that it carries a negative bias from the past. I think we should try to overcome that bias and have a bit of a more constructive view on it. But that also means being candid and transparent on the sort of risks that it entails and being assured that securitisation comes with significant risk transfer, especially in the case of synthetic securitisation structures.

Do you have a view on the merits of synthetic versus cash securitisation?

We are sort of agnostic, again provided that there are proper safeguards, risk management and governance in place. Synthetics have been growing quite substantially over the last years. Between 2022 and the end of 2025, transaction activity almost doubled, and the underlying exposures more than doubled in terms of what has been securitised. This suggests that European banks have been engaging in synthetic securitisation, which may have to do with the fact that these structures are more manageable for their capital management, and they see more demand from the investors that offer protection. But this entails crucial upfront risks. One is counterparty credit risk, and here the case of AIG from the financial crisis immediately comes to mind as an example of unfunded protection provided by insurance companies.

What do you make of the agreement in principle by European politicians for a bigger role for insurance companies in buying synthetic securitisations?

We don’t yet know the final text of this agreement in the trilogues. One of our concerns regards the proposed treatment of unfunded credit protection from insurance companies in synthetic securitisations qualifying for the STS label (simple, transparent standardised securitisations). Although the agreement contains positive aspects, the treatment of unfunded protection from insurance companies deviates from our recommendation. We consider this approach detached from the risks that might unfold in cases where the counterpart – the insurance company – is not able to continue honouring its protection.

What do you think of the preliminary agreement on risk weight floors?

We had expressed our concern in having risk-sensitive risk weight floors for senior positions. We prefer absolute floors, and, in particular, we would rather not lower risk weight floors on senior tranches below 7% (Basel II floor applicable before the great financial crisis). We understand that it will be below 7% for resilient STS securitisation. I think, in particular on the resilient STS, that we might be going too low in terms of the absolute floor and we have expressed our concern in the opinion.

Investors are increasing the use of insurance-backed structures to finance Synthetic Risk Transfers (SRTs), is that a problem?

We are now seeing a growing trend in terms of certain platforms being used by non-bank financial intermediaries to repackage synthetic securitisations. The insurance companies assume the credit risk from the credit-linked notes and the banks in turn provide the necessary financing for notes. This is a trend we started to see at the end of 2024, but usage started to grow in 2025.

Here we would clearly want to check and monitor whether these linkages might materialise or become amplified. For the moment, we don’t see a particular case of this sort of risks materialising, but we are monitoring the situation. The more you see the market grow, the more you see certain structures being used. The more complex the structure becomes, the more uncomfortable we are with these sorts of transactions. This market is also fairly opaque. If we see indications of securitisations becoming more complex, then we have an issue, and then of course these transactions will necessarily not qualify for simple and standard, which means that they will also be subject to a more intensive and comprehensive assessment, and hence will not qualify for the fast track procedure for SRTs.

Investors are increasingly using insurance to back loans for SRTs, is that an issue?

The more we see structures being designed that enhance complexity, the more we think there's a call for being more comprehensive in our broader assessment and the scrutiny that we have to do on transactions. The problem is always the same: the funding linkages to the banking system and whether there is significant risk transfer. Insurance-bank financing may come with the risk that these interlinkages build up. We not only see the risk of this flowing back to the banking system, but it also increases the complexity of the transaction. So yes, the sort of standardisation, simplicity and transparency that we would like to see in the transactions may be at stake.

We are seeing an increase of SRTs tied to data centres, does that pose a challenge from your perspective?

I don't have figures that I can share on the size of exposures that banks have to data centres. I would pose a slightly different question: how are the banks financing their AI exposures? Credit risk is a supervisory priority and this year we are carrying out a thematic review on underwriting standards. Like all large-scale infrastructure projects, banks need to ensure sound credit underwriting practices and ensure all risks are carefully assessed and weighted against the banks’ risk appetite. Data centre lending has specific dimensions. What really matters is which part of the cycle the banks are financing. With data centres, you have land acquisition, development, construction and the operational phase. Some parts are more related to real estate development while others have to do with energy sustainability. What we try to scrutinise is the credit underwriting practice regarding these specific elements and, in particular, how they are reflected in terms of credit risk assessment and pricing.

In addition to reviewing credit underwriting practices, the ECB regularly carries out analyses on potentially risky sectors to identify any concentrations or emerging risks which should be more carefully monitored or investigated.

Are there any takeaways you can offer on the balance of risk and opportunity that AI represents for European banks?

There is a large share of adoption of AI, both in terms of incorporating it in existing tools and trying to incorporate generative AI. Around 85% of the banks use generative AI already. The use cases tend to range from fraud detection, customer due diligence, notably know-your-customer documentation, transaction monitoring and credit scoring. There are different speeds and approaches in adoption. We are also seeing banks increasingly turning their attention to operational resilience and, in particular, cyber security. AI frontier models certainly change the paradigm in terms of cyber security because they shorten the time between a vulnerability being discovered and exploited. The biggest challenge for the banks today is really to put forward very swift and effective patching exercises and ensure the cyber hygiene of their systems. A more meaningful and substantial assessment can only be made after we get the banks' action plans (by the end of October). This is something that we clearly see as an area that banks need to get right in terms of governance and risk management against the backdrop of ensuring operational resilience.

Do AI agents that allow banking customers to continuously hop between the highest deposit offers spell a new era of hot money?

In addition to improving banks’ cost structures and operational efficiency, digitalisation has brought an increased probability of volatility of deposits. Still, ECB research does not show that the growing use of online banking or apps necessarily translated into more deposit volatility. Of course, this can be amplified in moments of stress, but this has not increased deposit volatility. The dimension you bring up might amplify this trend of deposit volatility with agentic AIs looking for the best return in terms of deposits. But the growing adoption of AI by banks may also offer the opportunity to counter this risk with commercial proposals offering more personalised product recommendations, dynamic pricing and targeted marketing by analysing customer profiles, financial behaviour and borrowing needs.

Do you think the European Commission's banking report next year will address what the ECB has been calling for in terms of a more integrated banking market?

I cannot judge the political dynamics around the integration proposals that are foreseen in the Commission's July communication. What I can say is that, for us, the measures they addressed were based on a correct diagnosis that Europe is missing a fully fledged integrated banking market which would allow for more enhanced cross-border banking activity and enable banks to gain scale. In the banking union lending is 80% domestic. In the corporate segment, cross-border lending accounts for only around 15%. Meanwhile, in the United States, to give you an example, almost 50% of corporate lending is cross-state. Deepening the integration of the Single Market in the banking union would allow banks, to reap the benefits by diversifying their revenue stream and their risks as well as to better allocate their capital and liquidity, especially those with a cross-border footprint.

You have previously said that liquidity waivers have rarely been applied in the banking union. You have about three years left on your term as an ECB Supervisory Board member, will you approve more of them in that time?

They are possible in the legislation. Of course, the fact that we don't have a complete banking union and that we don't have a credible backstop in terms of both deposit insurance and liquidity in resolution, and that today a Member State might find itself in a situation of having to reimburse deposits held at a foreign bank's local subsidiary, makes the case for granting liquidity waivers extremely difficult. That said, in a normally functioning and fully fledged banking union, liquidity would be managed from a group perspective and flow across borders. Today we have roughly €230 billion of high-quality liquid assets which are trapped at subsidiary level. If they were to move across borders, they would more effectively address the financing needs of the real economy. But of course this may only unfold if the appropriate safeguards are in place.

Are there parts of the Commission's initial communication on simplification that give you pause, for example a softening of the output floor?

The Commission addresses the need for a more holistic assessment of the capital requirements. In doing so, I would say that we should not lose sight of a holistic assessment of resilience and risks, when considering the various reform workstreams that are currently ongoing. It's important to assess the impact of various changes that you might make to Pillar 1 requirements, notably on the regulatory backstops such as the output floor or the Pillar 2 Guidance, where there are some proposals suggesting that it should only be applied to outlier banks that are below their minimum requirements in the adverse scenario of the stress test. Then there are the changes to the securitisation regime which will bring more capital relief for banks. That's why we need a holistic assessment of the broader resilience. We should avoid underestimating significant risks down the road. That is why regulatory backstops, including the output floor, should be maintained – to guard against risks being underestimated. We're just emphasising the prudence required in this debate on competitiveness – not blurring the line between a desirable simplification and streamlining the capital stack on the one hand and diminishing the level of resilience, which we might be seeing, on the other.