When the Loan Does Not Move but the Risk Does: The Asset Anatomy of Synthetic Risk Transfer

Synthetic risk transfer allows part of a loan’s credit risk to move while the lending relationship stays in place. The borrower may continue paying the same bank, which may retain the loan on its balance sheet and continue servicing it without assigning it to the investor. The resulting investment follows a defined layer of potential loss rather than ownership of the loan.

Economic exposure does not have to move with legal ownership.

On 26 August 2026, the Financial Times reported a sharp increase in insurer participation in significant risk transfer, or SRT, transactions. Citing the International Association of Credit Portfolio Managers, it reported that insurers provided approximately €4.7 billion of new unfunded SRT protection in 2025, up from €2.7 billion in 2024. Financial Times: Insurers pile into deals allowing banks to offload default risk

By the end of 2025, insurers had provided protection on approximately €10.9 billion of outstanding SRT tranches from transactions executed since 2019, linked to about €366 billion of underlying loans. Reporting on the same IACPM survey said that large-corporate and SME portfolios represented 58% of new insured business in 2025, while residential mortgages represented 20%. Claims Journal: Insurers Want a Bigger Slice of the Bank Risk Transfer Boom A synthetic SRT is not primarily a sale of loans. It is a transaction in which the bank identifies a reference portfolio and transfers a defined layer of potential credit losses to another party.

The loans can remain with the bank. The risk layer becomes a separate financial object. The specialist records and lifecycle models below describe review requirements and illustrations. They should not be read as claims that software has verified the rights or that every proposed field and relationship is an existing DaDepo feature.

Separate loan ownership from the layer of risk transferred

A traditional loan sale is intuitive.

A traditional securitisation can extend that chain:

Legal ownership of the underlying assets moves away from the originator, subject to the transaction structure and applicable law. Synthetic securitisation works differently. The International Association of Credit Portfolio Managers describes synthetic on-balance-sheet securitisation as a structure in which the securitised assets are not legally transferred out of the bank’s ownership. Instead, a third-party protection provider agrees to compensate the bank for specified losses on the assets. IACPM: Strengthening Financial Stability through Insurance-based Credit Risk Transfer The simplified chain becomes:

1. Borrowers
2. Loans remain with bank
3. Loans form reference portfolio
4. Defined loss tranche is transferred
5. Investor / insurer provides protection

The asset did not move. The risk did. The phrase reference portfolio is central. Suppose a bank has €10 billion of corporate loans. It may select €1 billion of eligible exposures for an SRT transaction. Those loans are identified as the reference portfolio. But the bank can remain: legal lender, contractual counterparty to the borrower, holder of the loan, servicer, collector of principal and interest, manager of amendments and waivers, and party responsible for workouts. The investor may never obtain ownership of any individual loan. Instead, the investor is exposed to losses calculated by reference to the identified pool. That means a data model cannot treat: “Reference portfolio” as synonymous with “Assets transferred to investor”.

The reference relationship is its own legal and economic link. Imagine a bank has the following portfolio: 1,000 corporate loans with total outstanding principal of €2 billion. The SRT may define a protected loss layer from 2% to 8% of the reference amount. The economic stack could look like:

The protection provider is not buying 6% of every loan. It is taking responsibility for losses that fall inside a defined horizontal slice of the portfolio loss distribution. That slice can itself have notional amount, attachment point, detachment point, premium or coupon, maturity, loss-allocation rules, replenishment rules, credit-event definitions, settlement mechanics, and termination provisions. This is the new financial object. The mechanics are easier to understand with an example. Assume a €1 billion reference portfolio. The bank retains losses up to 1%. A protection provider covers losses from 1% to 6%. Losses above 6% remain outside that protected tranche.

The attachment point is 1%. The detachment point is 6%. If portfolio losses are €5 million, the protection provider may owe nothing because losses have not reached the attachment point. If losses reach €30 million, €20 million may fall inside the protected layer, subject to the transaction documents. If losses reach €100 million, the protection provider’s maximum exposure may be limited to the €50 million tranche between €10 million and €60 million. The investor therefore does not have generic exposure to "the portfolio". The investor has exposure to a precisely defined loss interval.

Identify the economic audiences and protection providers

One underlying loan can now matter to several parties for different reasons. The borrower sees: “My loan with Bank A”. The bank sees: Loan asset, Customer relationship, Interest income, Credit exposure, and Reference exposure in SRT. The protection provider sees: “One contributor to losses in a reference portfolio”. Those are related views. They are not interchangeable. This matters for privacy as well. An SRT investor may need detailed reference-portfolio information to underwrite the risk without becoming the lender of record or receiving unrestricted access to every borrower document. Asset infrastructure therefore needs both linkage and access control. Synthetic SRT can be funded or unfunded.

In a funded structure, the protection provider typically places funds or collateral into the structure upfront. A simplified funded chain may look like:

1. Investor pays collateral / note subscription
2. Collateral supports protection obligation
3. Reference portfolio incurs losses
4. Losses reduce investor principal / collateral

The bank has protection backed by assets already placed into the transaction structure. An unfunded structure is different. The Bank of England describes unfunded SRT as a structure in which collateral is not provided upfront. Instead, the investor or insurer gives a credit guarantee against losses on the reference portfolio. Bank of England: Financial Stability Report, July 2026—Developments in significant risk transfer activity The chain becomes:

1. Bank pays protection premium
2. Protection provider promises to cover defined losses
3. Loss event occurs
4. Bank claims under protection contract
5. Provider pays if contractual conditions are met

Now another credit risk appears. The bank is exposed to the protection provider’s ability and willingness to pay. This is why the 2025 insurer data is significant. Insurers do not necessarily need to buy the bank’s loans. They can underwrite credit protection on a defined tranche. The Financial Times reported that the volume of new insurer-provided unfunded protection increased from approximately €2.7 billion in 2024 to €4.7 billion in 2025. That is a different investment relationship from buying a private-credit loan. The insurer may receive premium for taking a contingent obligation.

Its exposure depends on reference portfolio performance, attachment and detachment points, transaction maturity, credit-event definitions, loss-calculation rules, recovery mechanics, protection contract wording, and its own capital and risk management. The underlying loans remain assets of the bank. The insurer owns a contractual position in the risk-transfer structure, not the lending relationship itself. At year-end 2025, approximately €10.9 billion of outstanding insurer-protected SRT tranches were reportedly linked to around €366 billion of underlying loans. Those numbers should not be compared as though insurers had guaranteed every euro of those loans. They illustrate the opposite. SRT typically transfers a selected slice of portfolio losses.

A relatively small protection notional can relate to a much larger reference pool because the protected tranche is designed around a particular part of the loss distribution. This is why a system that stores only: a €366bn reference portfolio amount and €10.9bn of insurance protection without tranche structure can be misleading. The investor needs to know where the protection sits.

Follow borrower payments, losses and recoveries

The IACPM survey data reported in August showed that corporate and SME loans represented 58% of new insured SRT business in 2025. Residential mortgages represented another 20%. Those are very different underlying assets. Corporate loan risk may depend on borrower leverage, sector, collateral, covenants, refinancing, sponsor support, geographic exposure, and idiosyncratic default events. Mortgage risk may depend on borrower affordability, loan-to-value, property prices, interest rates, unemployment, arrears, cure behaviour, foreclosure timing, and recovery value. The SRT wrapper can be similar. The underlying credit data cannot be generic. A good platform therefore needs to separate transaction structure from asset-class schema. Consider a residential mortgage. The bank originates it. The homeowner continues paying the bank.

The bank retains legal ownership. The loan enters a reference pool. An investor provides protection on a mezzanine tranche of pool losses. Now the mortgage participates in two economic systems. The mortgage relationship links the homeowner and bank; the risk-transfer relationship links the bank and protection provider. The protection provider may never contact the homeowner. It may never receive a mortgage assignment. It may never appear on land-registry records. Yet deterioration in that mortgage can contribute to losses that reduce the investor’s SRT position. That is why loan ownership and credit-risk exposure must be different fields. In a true-sale securitisation, borrower payments may ultimately flow through the SPV structure to investors.

In synthetic SRT, the investor cash flow can be structurally different. The bank may continue receiving ordinary borrower principal and interest. The protection provider may receive: periodic protection premium, coupon on a credit-linked note, collateral investment return, or another agreed compensation. If defined losses occur, the protection provider may make a payment or suffer a reduction in principal. There are therefore two cash-flow systems: “Borrower -> Bank” and “Bank <-> Protection provider”. They are economically linked through credit performance. They are not the same payment waterfall. The value of an SRT position depends on how losses are recognised.

A reference borrower may: miss a payment, enter arrears, default, restructure, enter insolvency, have collateral enforced, cure, make a recovery payment, or be written off. The transaction documents define which events matter and how losses enter the protected tranche. That can involve concepts such as credit event, realised loss, expected loss where permitted by structure, write-down, recovery, loss adjustment, allocation date, protection payment, and later reversal or recovery allocation. A portfolio tape showing only current principal balances is therefore insufficient. The SRT needs an event history. Suppose a borrower defaults and the bank recognises a €5 million loss. The protection provider compensates the bank for the part allocated to the covered tranche.

Two years later, the bank recovers €2 million through enforcement. Who benefits from that recovery? The answer depends on the contract. The protection mechanism may require some recovered amounts to be credited back or otherwise reflected in the transaction. That means the SRT lifecycle may continue long after the initial default event. A reliable record should connect:

1. Borrower default
2. Loss calculation
3. Tranche allocation
4. Protection claim
5. Protection payment
6. Later recovery
7. Recovery allocation

Without that chain, investor performance can be difficult to reconstruct.

Keep portfolio eligibility separate from capital relief

Reference portfolios change. Loans amortise. Loans mature. Borrowers refinance. Facilities are drawn. Some transactions allow replenishment or substitution during defined periods. Borrowers migrate between rating categories. Collateral values change. A mortgage goes into arrears. A corporate borrower breaches a covenant. A loan restructures. The SRT investor therefore needs to understand the portfolio as of each reporting date. Useful fields include opening reference balance, current reference balance, additions, removals, amortisation, defaulted exposure, cumulative loss, recovery, concentration, rating migration, and remaining tranche notional. An Asset Passport should preserve the difference between what was true at closing and what is true now. An SRT portfolio may include eligibility criteria.

For example, the transaction may define limits around: jurisdiction, borrower type, credit quality, maturity, arrears, loan type, collateral, sector, concentration, currency, or origination date. A loan can be included because it satisfied those rules at a particular time. Later events may affect whether it remains within the portfolio or how it is treated. The useful record is not simply: “Eligible: Yes”. It is: the eligibility rule, observed loan attribute, as-of date, result, any exception or waiver, and reviewer or source. That makes the inclusion decision auditable. Banks use SRT partly because transferring significant credit risk can reduce regulatory capital requirements associated with a portfolio when the relevant rules and supervisory requirements are satisfied.

This is economically important. It should not be confused with the asset itself. The chain is:

1. Reference portfolio
2. Risk-transfer structure
3. Supervisory / regulatory assessment
4. Capital treatment

The capital benefit depends on the applicable prudential regime and the transaction satisfying relevant conditions. A protection contract does not automatically create a fixed amount of capital relief. A regulatory-capital field should therefore have jurisdiction, regulatory framework, calculation date, bank model or approach, relevant approval or assessment status, capital impact, and source. It should not be treated as a permanent characteristic of the underlying loans.

Examine counterparty, servicing and consent dependencies

The Bank of England’s July 2026 Financial Stability Report highlighted the distinction between funded and unfunded SRT. In funded transactions, collateral provided upfront can substantially mitigate counterparty credit risk for the bank. In unfunded transactions, the bank relies on a guarantee. The protection provider itself therefore becomes a credit exposure. The PRA has identified risks including late payment of protection; non-payment; downgrade of the protection provider; and potential loss of eligibility of the provider for prudential recognition. Bank of England: SS9/13—Securitisation: Significant Risk Transfer That means a complete SRT record needs data on both sides: Reference credit risk, and Protection provider credit risk. The transaction reduces one risk by creating exposure to another contractual counterparty.

The underlying bank usually remains responsible for the borrower relationship and monitoring. That is operationally efficient. The bank knows the borrower. It originated the loan. It has systems for payment collection, covenant monitoring and workouts. But economic risk-sharing creates an incentive question. In March 2026, an ECB working paper using euro-area transaction-level data examined how banks behave after synthetic risk transfer. The authors reported evidence that banks reduce monitoring effort relative to other banks lending to the same firm, with larger reductions where a greater share of exposure has been synthetically transferred. They also highlighted growing links between banks and non-bank SRT investors. European Central Bank: Synthetic, but how much risk transfer?

The paper is research, not a conclusion that all SRT weakens monitoring. But it illustrates why servicing behaviour remains relevant to investors even when the bank retains the loans. The risk buyer depends partly on the risk seller continuing to manage the portfolio well. A synthetic investor may need information on arrears management, covenant waivers, restructurings, collateral releases, extensions, refinancing, borrower downgrades, forbearance, enforcement, recoveries, and write-offs. These actions can change loss outcomes. A reference portfolio is not passive merely because ownership remains unchanged. The bank’s management of the assets continues to affect the investor’s risk. That makes servicing governance part of the SRT Asset Passport.

Loan transfers often raise contractual questions around: borrower consent, notification, assignment restrictions, confidentiality, security transfer, and registration. A synthetic SRT can avoid some of the operational consequences of a loan sale because the lender of record may remain the same. That does not mean there are no legal or confidentiality issues. Portfolio data may be shared with investors. Credit protection documents need enforceability. Regulatory and disclosure requirements may apply. But the distinction is important: “Transfer of loan” and “Transfer of risk referencing loan” are different legal events. A platform should never infer one from the other.

A proposed evidence model for the loans and protection contract

SRT highlights a broader DaDepo design principle. A financing ecosystem may contain assets built on top of other assets. The underlying loan needs a record. The reference portfolio needs a record. The protected tranche needs a record. The guarantee, derivative, credit-linked note or insurance contract implementing the protection needs a record. Those records should be linked rather than collapsed. A useful structure may look like:

1. Loan Asset Passports
2. Reference Portfolio Passport
3. Risk Tranche Passport
4. Protection Contract / Instrument Passport
5. Investor Exposure

Each object has its own lifecycle. SRT is executed at portfolio level. That does not eliminate loan-level data. Portfolio losses emerge from individual credit exposures. A reviewer may need to trace a reported loss back to borrower, facility, current balance, default event, collateral, recovery, and transaction treatment. The portfolio record should therefore aggregate rather than replace asset-level identity. This is especially important when portfolios contain thousands of mortgages or SME loans. A summary can show concentration. Only the underlying asset record can explain a specific exception. Suppose a system contains:

What does "Investor" mean? Does Fund B own the loan? Own a participation? Own a credit-linked note? Sell protection under a derivative? Provide an insurance guarantee? Hold a funded SRT tranche? Those positions can have very different rights. A better model separates legal owner, lender of record, servicer, participant, secured creditor, protection buyer, protection seller, tranche investor, noteholder, insurer, guarantor, and beneficiary. The words matter because the rights differ. For DaDepo, the practical response is not to invent one asset type called SRT and place everything inside it. The structure should preserve at least four layers.

1. Underlying loans

Borrower, lender, facility identifier, loan type, origination date, maturity, currency, original balance, current balance, interest terms, collateral, guarantees, payment status, internal rating, default status, recovery status, and source documents.

2. Reference portfolio

Portfolio identifier, originator, reporting date, eligible assets, inclusion date, removal date, reference balance, concentration limits, replenishment rules, defaulted balance, cumulative losses, recoveries, and reporting methodology.

3. Protected tranche

Tranche notional, attachment point, detachment point, initial thickness, current thickness, maturity, amortisation, loss allocation, premium or coupon, funded or unfunded status, and remaining protection amount.

4. Protection contract or instrument

Protection buyer, protection provider, legal instrument, guarantor or insurer where relevant, collateral where funded, credit-event definition, loss-calculation method, claim procedure, payment timing, recovery treatment, termination events, transfer restrictions, governing law, and current status.

Provenance across all four layers

Source document, source system, calculation source, reporting period, user-entered data, extracted data, reviewed data, external confirmation, reviewer, version, and last updated date. That model can answer a question that one generic loan record cannot: Which rights stayed with the bank and which risks moved to the investor?

AI support and the bounded DaDepo contribution

AI-assisted preparation

The analytical uses discussed here are possibilities for document review, subject to source access and human judgement. They are not a claim that DaDepo currently performs every proposed check or reaches the resulting legal or financial conclusions. Large SRT portfolios are obvious candidates for AI-assisted data preparation. AI can help classify loan documents, reconcile loan tapes, identify borrower entities, extract balances, map collateral, detect missing fields, compare reporting periods, identify loans added or removed, flag duplicate references, identify covenant events, detect arrears, connect default notices, reconcile recovery records, and prepare loss-calculation evidence. It can also help compare transaction documents with portfolio reporting. For example:

That is useful. AI should not independently determine: that legal SRT requirements are satisfied; that regulatory capital relief is available; that a credit event legally occurred; that a protection claim is payable; that a loss calculation is contractually final; that a protection provider remains eligible; that a loan is enforceable; that collateral is correctly valued; that servicing was prudent; or that the SRT investment is suitable. The system can organise evidence and calculations. The authoritative decision remains with the relevant bank, investor, insurer, calculation agent, legal adviser, auditor, regulator or other responsible party.

DaDepo can help prepare a private document-backed asset draft and retain source-linked information for human review. The private Parties and contractual roles table can record source-described party, contract and role rows with document and clause references, optional source dates, and pending, confirmed or rejected review decisions. The same party can appear in different contracts and roles. Where current compatible completed analysis is available, AI may suggest pending role rows; a user reviews and confirms or rejects them. These records do not establish legal title, priority, enforceability, the validity of a reported exposure or the outcome of a dispute.

The specialist information model for synthetic risk transfer is a proposed information requirement. A complete relationship graph, automatic exposure reconciliation and automatic cross-contract decisions are not delivered capabilities established by this article. DaDepo’s role is not to become an SRT arranger. The opportunity is information architecture. Synthetic risk transfer creates exactly the type of multi-layer private asset that becomes difficult to reconstruct from disconnected documents and spreadsheets. A proposed specialist workflow would need to connect: underlying loans, borrower evidence, reference portfolio membership, reporting periods, tranche terms, protection agreements, counterparties, loss events, recoveries, claims, payments, and provenance.

If implemented, this proposed model could support: bank credit-portfolio management; investor diligence; insurer underwriting; private-credit analysis; mortgage portfolio review; transaction monitoring; audit; servicing oversight; calculation-agent workflows; and later transfer or refinancing of the protection position where legally and contractually permitted. The Asset Passport does not create the risk transfer. The proposed model aims to make the risk-transfer structure understandable.

Creating or reviewing an SRT-related Asset Passport does not mean that DaDepo has originated or purchased the underlying loans; verified every loan balance; assessed borrower creditworthiness; confirmed collateral value; determined that a reference exposure is eligible; calculated regulatory capital requirements; determined that significant risk transfer has occurred; obtained supervisory approval; issued a credit-linked note; sold or purchased credit protection; provided an insurance policy or guarantee; determined that a credit event has occurred; calculated a legally binding loss; adjudicated a protection claim; guaranteed payment by a protection provider; valued an SRT tranche; rated a securitisation; recommended an investment; provided banking, insurance, reinsurance, securitisation or derivatives services; or provided legal, regulatory, investment, tax, accounting, credit-rating, underwriting or valuation advice.

Important: DaDepo provides technology and information tools. It does not provide legal, financial, investment, tax, accounting, banking, insurance, reinsurance, securitisation, derivatives, regulatory-capital, credit-rating, underwriting, settlement or valuation advice or services unless a specific service is expressly identified and lawfully provided. SRT structures, prudential treatment, protection eligibility, loss recognition and investor rights vary by jurisdiction, transaction and facts. Users should review the governing documents and obtain appropriate professional advice.

Questions for a synthetic-risk-transfer review

Before an SRT reference portfolio or protection position is presented to an investor, insurer, reviewer or regulator, ask:

1. Underlying assets: What exact loans or exposures form the reference portfolio?
2. Ownership: Who legally owns each underlying loan?
3. Lender of record: Who retains the borrower relationship?
4. Servicing: Who monitors, collects and works out the loans?
5. Reference balance: What is the current portfolio balance and as of what date?
6. Eligibility: Why is each exposure included in the reference portfolio?
7. Portfolio changes: Which loans have been added, removed, repaid or substituted?
8. Risk layer: What are the attachment and detachment points?
9. Protection amount: What is the initial and current protected tranche notional?
10. Instrument: Is protection provided by guarantee, insurance, derivative, credit-linked note or another structure?
11. Funding: Is the protection funded or unfunded?
12. Protection provider: Who bears the transferred risk and what is its current credit standing?
13. Credit events: Which borrower events can create a protection claim?
14. Loss calculation: How is a portfolio loss calculated and allocated to the tranche?
15. Recovery: How are later recoveries treated after a protection payment?
16. Cash flows: Which payments are borrower cash flows and which are SRT premium, coupon or protection payments?
17. Capital treatment: What regulatory framework and assessment support any claimed capital relief?
18. Transferability: May the protection position or instrument itself be transferred?
19. Monitoring: Are material servicing decisions and portfolio changes visible to the risk holder?
20. Provenance: Can every reported exposure, loss, recovery and tranche calculation be traced to its source and reporting period?

If the system cannot distinguish who owns the loan from who bears the loss, it does not yet understand the transaction.

Private credit beyond direct loan ownership

Private credit is often described as capital moving from banks to non-bank lenders. Synthetic risk transfer shows another direction. Non-bank capital can take credit exposure without originating the loan. Without servicing the borrower. Without purchasing the loan. Without appearing as lender of record. The investment can exist one layer above the asset. That creates a broader private-credit map: direct loan ownership, loan participation, true-sale securitisation, funded synthetic risk tranches, unfunded credit protection and insurance-based SRT. All provide exposure to credit. They do not provide the same legal rights. The growth of insurer participation makes this distinction more important, not less. Asset infrastructure designed only around ownership transfer will miss a growing part of the market.

The future system needs to model: Legal asset, Reference relationship, Economic exposure, Transferred risk, Contractual protection, Lifecycle, and Provenance. A loan can stay on the same balance sheet for its entire life. The risk around it can move several times. That risk is becoming an asset in its own right.

Further reading

• Financial Times: Insurers pile into deals allowing banks to offload default risk
• Claims Journal: Insurers Want a Bigger Slice of the Bank Risk Transfer Boom
• International Association of Credit Portfolio Managers: Global SRT Insurance Survey
• IACPM: Strengthening Financial Stability through Insurance-based Credit Risk Transfer
• IACPM: Risk Sharing is Becoming Mainstream
• Bank of England: Financial Stability Report—July 2026
• Bank of England: SS9/13—Securitisation: Significant Risk Transfer
• European Central Bank: Synthetic, but how much risk transfer?