CR 23 – An Introduction to Securitisation

A bank holding ten thousand car loans owns a stream of monthly cash that nobody outside the bank can easily buy. Securitisation changes that. The bank sells the loans to a separate company created only to hold them, and that company pays for the purchase by issuing notes to investors. The interest and principal the borrowers pay on the car loans is then the money used to pay interest and principal on the notes. An asset that was illiquid and sat on one balance sheet becomes a rated, tradable security that any qualifying investor can hold.

The notes produced this way are called asset-backed securities (ABS). They are debt instruments whose coupons, often floating rate, are serviced from a pool of loans rather than from the profits of an operating business. By convention, notes backed by residential or commercial mortgages are grouped separately as mortgage-backed securities (MBS), and the label ABS is kept for everything else, such as credit card balances, auto loans, equipment leases and trade receivables. Modern structuring techniques can build a bond out of almost any predictable cash flow, which is why the asset list has grown so long.

Sundaresan (1997) describes the technique as a framework in which illiquid assets of a corporation or financial institution are turned into a package of securities through packaging, credit enhancement, liquidity enhancement and structuring. Each of those four words matters later in this lesson: packaging is the pooling of many loans, credit enhancement is the protection given to the notes, liquidity enhancement is the creation of something tradable, and structuring is the division of the notes into ranked classes.

Cash securitisation versus synthetic securitisation

The technique began with US mortgage banks in the 1970s as a way of funding their lending, and it then spread to credit card payments, equipment leasing receivables and other assets. Banks also came to use it as an asset and liability management tool, a way of controlling balance sheet risk. In a cash flow securitisation the assets are actually sold to a separate legal entity, the special purpose vehicle. A synthetic securitisation, whose market dates only from 1997, transfers the credit risk of a pool through credit derivatives while the assets themselves stay where they are. The two aim at the same broad result: bringing those who need capital or want to shed risk together with those who can supply capital or want to take on risk, without a bank standing in the middle. Institutional investors buy both kinds of product, and while they may not care why a deal was originated, they care a great deal about the name and quality of the institution behind it.

The rest of this lesson deals with cash securitisation, since understanding the cash version is the foundation for every synthetic variant.

The special purpose vehicle (SPV), also called a special purpose entity (SPE) or special purpose company (SPC), is the legal shell through which a securitisation is carried out. Assets such as mortgages or credit card debt, or synthetic exposures such as credit derivatives, are transferred into it, and the SPV reshapes their risk and reward into notes that a wider range of investors can buy. Its whole reason to exist is to make risks tradable that would otherwise have no secondary market.

What makes the SPV useful is that it has no legal link to the originator. It is bankruptcy-remote: if the sponsoring bank runs into trouble or is declared bankrupt, the SPV and the notes it has issued are untouched, because the assets now belong to the SPV and not to the bank. The flip side is that investors are exposed only to the assets inside the vehicle. If those assets deteriorate and the notes are downgraded, the noteholders have no claim against the originator. Where the currency or interest basis of the assets differs from that of the notes, the SPV enters currency or interest rate swaps with a bank, and it then carries counterparty risk to that bank as well.

Where SPVs are located and how they are set up

Tax drives the choice of domicile. The sponsor wants every cash flow received and paid by the vehicle to attract little or no tax, including withholding tax on the coupons paid to noteholders. So SPVs are often incorporated in offshore centres such as Jersey or the Cayman Islands, or in places that have written SPV-friendly legislation, such as Dublin or the Netherlands. Operating costs, legal requirements and investor preferences also count; some investors in European Union countries will only buy notes from an SPV domiciled in the EU, which rules out many offshore locations.

Legally an SPV is either a company or a trust. The trust form is more common in the US, where the vehicle’s interests are represented by a trustee, usually the agency services department of a large bank or a specialist trust company. In the Euromarkets SPVs are more often incorporated as companies.

Common uses of an SPV

The most frequent use is the cash flow securitisation itself: the sponsor sells assets off its balance sheet to the SPV, the SPV funds the purchase by issuing notes, and the income from the assets services those notes. Untradable assets such as residential mortgages become tradable ones, and the originator’s balance sheet is freed. SPVs are used in several other ways too:

• Repackaging an asset swap. The SPV buys an asset swap package (a fixed coupon bond swapped into floating rate with a swap counterparty) and issues notes against it, so investors who could not deal in asset swaps directly gain the exposure indirectly.
• Converting currency. A currency swap inside the SPV turns the currency of the underlying assets into one investors prefer.
• Issuing credit-linked notes. Unlike a credit-linked note issued directly by a bank, one issued by an SPV carries no credit link to the sponsor; its performance depends on the assets sold into the vehicle. In another version, investors choose collateral that the SPV holds, the SPV sells credit protection to a swap counterparty, and after a credit event the collateral is sold to meet the SPV’s liabilities.
• Making illiquid assets liquid. Trade receivables, equipment lease receivables and even unusual cash flows such as museum entry fees cannot be traded in any form, but securitisation can turn them into tradable notes.

Banks securitise to unlock value tied up in their balance sheet assets, typically residential mortgages, corporate loans and retail credit such as card debt. Three immediate effects make this attractive. If income stays roughly the same while total assets shrink, return on equity rises. Less capital is needed to support the smaller balance sheet, which saves cost or frees capital for more profitable business. And funding can be cheaper, because the interest paid on the asset-backed notes is often well below the rate earned on the underlying loans, leaving a cash surplus for the originator. These effects group into three motives: funding, balance sheet capital management, and risk management.

Funding motives

Securitisation lets a bank grow its assets quickly, diversify its funding and lower its cost, and reduce maturity mismatches. The ABS market is large, estimated at $1,000 billion invested worldwide in 2003, so a bank with access to it can expand its loan book faster than deposits and conventional borrowing would allow. Northern Rock, a former UK building society turned bank, used securitisation to fund a growing share of the UK mortgage market; it came to depend on it too heavily at the expense of retail deposits, with disastrous results when markets seized up in 2007.

Diversification matters because relying on one or a few funding sources is dangerous when markets are stressed, and banks try to balance retail, interbank and wholesale money. Cost falls because securitisation separates the credit rating of the notes from the credit rating of the bank. Most notes issued by SPVs are rated higher than the originator’s own bonds, so even though ABS trade in a less liquid secondary market than corporate bonds and pay some extra yield for that, the all-in cost is often still lower.

The maturity point concerns asset and liability management. A bank naturally funds long assets, such as 25-year mortgages, with short liabilities such as deposits and interbank borrowing. Selling the mortgages to an SPV that issues notes of matching economic maturity removes that gap for the assets involved.

Capital management and risk transfer

On the capital side securitisation offers regulatory capital relief, economic capital relief and a more diverse capital base. Under Basel I, in force since 1988 and effective from 1992, a bank had to hold at least $8 of capital for every $100 of risk-weighted assets, and the risk weights were crude: apart from mortgages, customer loans were weighted at 100% whatever the borrower’s rating or collateral. That rigidity, addressed by Basel II from 2008, was a strong driver of securitisation. An SPV is not a bank, so it holds only the capital the economics of its assets demand, which is always far below the 8% a bank needs. The originator does not get full relief, because it usually keeps the first-loss piece, but its capital charge falls sharply. In this sense securitisation works as a substitute for raising Tier 1 equity, preferred shares or perpetual step-up notes, and the smaller capital base lifts return on equity, which shareholders like. Accounting consolidation rules, however, make full capital relief hard to obtain in many deals, so the other benefits carry more weight.

For risk management, selling assets to the SPV cuts the bank’s credit exposure to them sharply and, if no first-loss piece is retained, removes it entirely. Banks can also securitise non-performing loans. That removes both the credit risk and the negative signal those loans send, frees capital, and still leaves an upside: if some loans recover or defaulted assets yield recoveries, surplus profit made by the SPV flows back to the originator.

Investor demand for ABS has been strong since the market began, for three main reasons. Investors can spread their holdings across sectors, they can reach risk and return profiles that may be better than what is otherwise available, and they can enter sectors such as consumer credit that would otherwise be closed to them.

Tailoring is the most distinctive advantage. If the market is short of assets with a particular rating, securitisation can manufacture them by cutting a pool into tranches of the required quality. Securitised notes have also often performed better, in credit terms, than corporate bonds with the same rating and maturity. That can look odd, since two AA-rated bonds should behave alike, but the explanation is structural: the originator usually holds the first-loss piece, which absorbs early losses before any rated note is touched and gives the originator a reason to select and service the pool carefully.

Diversification is the other argument. An investor who puts $100 million into one AA-rated corporate bond is exposed to event risk at that single issuer, whereas the same $100 million in an ABS spreads exposure over, say, 100 pooled assets, lowering concentration risk. That was the theory. The 2007 to 2008 crash showed that in some cases pooling actually increased concentration, because many pools were filled with the same kind of fragile loan.

The originator is the bank or company that owns the loans or receivables to be sold. The issuer is normally an SPV set up for the deal and often domiciled offshore, and it buys the assets from the originator. Keeping the pool inside a separate vehicle means an insolvency of the originator cannot reach the transferred assets, so the originator’s own rating becomes almost irrelevant to noteholders. Credit enhancement, sometimes including a third-party guarantee, then lifts the notes to investment grade and often to AAA.

The structure is designed so that the notes, which form the SPV’s liabilities, cost less than the asset side yields. That spread is the source of the originator’s cheaper funding, a benefit that matters most to institutions with weaker credit ratings that would pay much more for unsecured borrowing.

Mechanics of a true sale

The heart of the process is a true sale: the assets leave the originator’s balance sheet and move onto the SPV’s. Because the transfer must be recognised as a genuine legal sale, the originator takes legal advice on it. The steps are:

• investigating the assets in a due diligence review of their quality and outlook;
• establishing the SPV and moving the assets into it;
• underwriting the loans for credit quality and servicing;
• deciding the note structure, including how many tranches to issue, to suit both the originator and investors;
• having the notes rated by one or more rating agencies;
• placing the notes in the capital markets.

The rating agencies assess what kind of assets are in the pool, how good they are, and any enhancement that raises their credit quality, such as over-collateralisation or a liquidity facility from a bank. A major decision for the originator is which investment bank to mandate as underwriter; it structures the deal and places the notes, and is chosen on fees, distribution strength and its record with the asset class.

There are two main SPV structures, amortising and revolving, and a third, the master trust, used by issuers who come to market often. The choice depends mostly on how long the underlying assets last and how fast they repay.

Amortising (pass-through) structures

An amortising structure passes principal and interest to investors on each coupon date throughout the life of the notes, so the outstanding balance falls steadily, with the interest portion of each payment shrinking as principal is returned. Because investors do not know exactly when principal will arrive, these notes are priced and traded on their expected maturity and their weighted average life (WAL), the time-weighted period over which principal remains outstanding. WAL depends on a prepayment assumption: faster prepayment returns principal sooner and shortens it, slower prepayment lengthens it. Pass-through structures are standard for residential and commercial MBS and for consumer loan ABS.

Revolving structures

A revolving structure recycles principal. During the revolving period, principal collected from the pool is not paid to investors but used to buy new receivables that meet set eligibility criteria, so the pool stays the same size. This suits short assets that repay quickly, such as credit card balances and auto loans; without revolving, such a pool would shrink almost at once. When the amortisation period begins, investors are repaid in one of two ways. Under controlled amortisation principal is returned in a series of equal instalments. Under a soft bullet principal is collected in a separate account and paid out in a single lump sum on the expected maturity date.

Master trust

Issuers who come to market repeatedly under US and UK law rely on a master trust, which lets many securitisations be issued from one SPV. The originator transfers assets into the master trust, and new series of notes are issued from the shared pool whenever investor demand allows. MBS and credit card ABS originators are the main users. For credit cards, the master trust replaced the stand-alone trust as the preferred vehicle from 1991.

The notes an SPV issues are not identical. They are cut into tranches, each designed to carry a defined part of the pool’s risk, and each receives a different rating. The senior tranche is usually rated AAA. Tranches below it are rated progressively lower and are protected by a degree of over-collateralisation, which means they can absorb some losses before investors in them lose money.

The most junior tranche is either unrated or carries the lowest rating. It is the first-loss piece, so called because any loss in the asset pool hits it before reaching any other note. It is also known as the equity piece or equity note, despite being a bond in legal form, because like shareholders’ equity it takes losses first and receives whatever is left over once the senior claims are paid. The originator usually keeps it. That retention protects the senior investors, ties the originator’s fortunes to the pool’s performance, and, as the capital example later shows, determines how much capital relief the originator actually gets.

Ordering losses from the bottom up is what lets a pool of, say, BBB-quality loans support a large AAA tranche: the senior notes are paid in full unless losses exceed everything beneath them. The thickness of the junior layers is what the rating agencies size when they decide whether the senior notes deserve AAA.

Credit enhancement is the set of protections built into an ABS or MBS issue so that the notes achieve the rating investors require. The weaker the assets, the more enhancement is needed. Five methods are common, and most deals combine several of them.

• Over-collateralisation. The face value of the assets in the pool exceeds the face value of the notes issued against it, so the pool can lose the difference before noteholders are short of principal.
• Pool insurance. A composite insurance company writes a policy covering principal losses on the collateral. The insurer’s claims-paying rating is an important input into the rating of the notes, since the cover is worth only as much as the insurer’s ability to pay.
• Senior and junior note classes. A class B note is subordinated to the class A notes: its share of cash flows is paid only after the senior rights are met. Class B receives no principal until rating agency conditions are satisfied, usually a period of satisfactory pool performance, and in many deals not until the senior classes are fully repaid.
• Margin step-up. The coupon rises at a set date, usually a call date. The issuer is not obliged to call the notes, but the step-up makes keeping them outstanding expensive, so investors can reasonably expect the issuer to refinance and exercise the call.
• Excess spread. The pool earns more than the notes cost. Each month the difference between the asset yield and the note coupons first pays expenses and covers losses. Any surplus goes into a reserve account to cover future losses or, if not needed, eventually returns to the originator. While it sits in the reserve, it protects investors.

The methods differ in who provides the protection. Over-collateralisation, subordination and excess spread come from inside the deal, from the pool and the structure itself. Pool insurance, and the financial guarantees described in the airline case study later, come from an outside party whose own credit then matters. The margin step-up is different again: it protects investors against extension risk, the chance that the notes stay outstanding longer than expected, rather than against default losses.

Every securitisation pays its cash in a fixed order of priority, known as the cash waterfall. Interest collected from the collateral pool flows down a ladder of claims, and a lower rung is paid only once every rung above it is fully satisfied. An independent agent runs coverage tests at set points to confirm that the vehicle has enough cash to meet its obligations. If a test fails, cash that would have gone further down the ladder is redirected to repaying the notes, starting with the most senior.

This ordering is a form of credit enhancement in its own right. When the pool performs well, the junior classes and the equity piece collect their interest and the residual. When it starts to deteriorate, the tests trip and the cash that would have rewarded the junior holders instead shrinks the senior notes, so the protection beneath the senior class grows exactly when it is needed. Principal on Class B is paid only once Class A has been redeemed in full, and principal on Class C only after Class B, with any residual going to the subordinated notes.

Before a securitisation, the loans sit on the originating bank’s balance sheet and the bank holds capital against all of them. Afterwards the selected assets have left the balance sheet, but the bank usually keeps the first-loss piece. Under the original Basel I treatment the retained first-loss amount comes straight off the bank’s total capital, so the bank still bears capital equal to that piece, but nothing more.

Many banks run above the regulatory minimum. The minimum is 8%, while a target of 12% to 15% or more is normal practice, so the relief from a securitisation should be measured against the ratio the bank actually targets.

A hypothetical deal shows the questions an arranging bank works through. A fictitious airline, ABC Airways plc, wants to securitise its future ticket receivables, and the deal is arranged by an equally fictitious XYZ Securities. This is far from a plain vanilla transaction; it shows how far the technique was stretched toward exotic assets before 2007. After the 2007 to 2008 crisis, investors turned risk averse, deals of this kind stopped closing, and issuance returned to conventional asset classes.

Due diligence and marketing

XYZ Securities studies the airline’s performance over the last five years and projects forward the airline’s passenger sales, ticket sales and credit card receivables, together with where in the world its tickets are sold. What is being securitised is the future flow of credit card payments for tickets. This is riskier than a mortgage pool because airline earnings have historically been far more volatile than those of mortgage lenders.

All existing and future card receivables from ticket sales are transferred to an SPV, and the bank’s syndication desk aims the notes at institutional investors across Europe. An indicative price is set first to test demand. In November 2002 the indicative spreads over 3-month LIBOR were 70 to 80 bps for the AA note, 120 to 130 bps for the A note and 260 to 270 bps for the BBB note, levels that look very low by post-2007 standards. The exact spreads are fixed during the roadshow, and the pricing is benchmarked against recent deals in similar assets and against unsecured spreads of comparable issuers.

Deal structure and guarantee

ABC Airways sells its future ticket receivables to an offshore SPV, Airways No 1 Ltd, which issues notes to fund the purchase. The SPV pledges its rights over the receivables to a Security Trustee, a fiduciary acting for the bondholders, which collects funds as the SPV receives them. Bondholders are paid interest and principal quarterly in order of note priority, and if the deal defaults the trustee acts to protect them. The arranger also decides whether to ask a mono-line insurer to wrap the deal, guaranteeing the SPV’s obligations in return for a fee.

Financial modelling and the DSCR

The arranger builds a cash flow model to size the notes, using historical sales, seasonality and card cash flows, plus assumptions on growth, inflation and tax, run across several scenarios. A key output is the cover that the asset cash flow provides over the debt payments, known as the debt service coverage ratio (DSCR). A more conservative DSCR gives investors more comfort. For a residential mortgage deal a DSCR of about 2.5 to 3.0 may be acceptable, but for airline ticket receivables it would rarely be set below 4.0. The model then finds the largest note issue the assets can support while keeping the DSCR at that floor.

Credit rating and after issue

A rating from Moody’s, Fitch or Standard & Poor’s makes the notes far easier to place. The agencies blend qualitative and quantitative analysis tailored to the asset class. For this deal they would examine the airline’s corporate credit quality (its financial record, liquidity and debt, whether it is state-owned, conditions in the economy and the airline industry, its safety record and fleet age), its market share and competition, regulatory changes that could hit cash flows, the legal structure of the SPV and the asset transfer, and the cash flow analysis. If needed, the arranger reshapes the deal to reach the target ratings. From start to close the process can take three to 12 months or more. After issue the arranger steps away, while agency services such as the paying agent, cash manager and custodian continue until the notes are repaid.

The MBS market began when US government-chartered mortgage agencies issued pass-through securities backed by residential mortgages to make home loans cheaper. It grew because investors gained a liquid instrument and lenders could shift interest rate risk off their books. US finance companies then applied the same methods to non-mortgage receivables, creating the ABS market. MBS investors had enjoyed implicit Treasury support through the Ginnie Mae issues, whereas ABS gave investors a different portfolio dynamic and access to a wider range of credit classes.

Between 2002 and 2007, low interest rates and frequent corporate downgrades made rating-resilient ABS and MBS attractive, and they traded at yields that compared well with similarly rated unsecured debt. European securitisation issuance exceeded €157.7 billion in 2003. In the US, auto-loan and credit card ABS lead alongside agency MBS, while in Europe residential mortgages dominate: RMBS made up over 55% of all European securitisation issuance and over 90% of European MBS in 2003, helped by a strong UK housing market. Commercial MBS benefited from favourable insolvency rules and from the euro, which removed currency concerns.

How ABS collateral differs from CDO collateral

ABS performance depends heavily on consumer credit, so ABS structures typically include triggers that accelerate amortisation and reserve accounts that cover interest shortfalls. Their basic structure is the same as that of a CDO; the differences come from the collateral and the issuer’s motive. A CDO portfolio might hold 100 to 200 loans, while an ABS pool often contains thousands of obligors, which gives it the diversity a consumer pool needs.

Auto loans

Auto loans were among the first assets securitised. Investors liked their high quality and the fact that the car is a tangible asset that can be sold readily if the borrower defaults. A car is regarded as an essential purchase, and the loans are short, typically 3 to 5 years, which leaves little reason to refinance, so there is no real prepayment culture. Prepayment speeds are very stable and losses relatively low, especially among prime borrowers.

Two tools dominate auto ABS analysis. Loss curves show the expected cumulative loss over the life of a pool; set against actual losses they measure how the pool is performing, and the loss forecasts are especially useful to buyers of subordinated classes. Prime pools tend to show losses spread evenly, while non-prime and sub-prime pools recognise losses earlier and so have steeper curves; in both cases losses usually decline in the later years. The absolute prepayment speed (ABS), first developed by Credit Suisse First Boston, compares each period’s prepayments with the pool balance. Like every prepayment measure, it indicates the expected maturity of the notes and, in effect, the value of the call option embedded in them.

Credit card pools have no fixed loan term, and a single cardholder’s balance often lasts no more than six months. The structure must therefore control how fast investors are repaid, either through scheduled amortisation or through a revolving period in which collections buy new receivables. For specialist US card banks, ABS became the main way to fund their large volume of unsecured consumer lending, and since 1991 the master trust has been the standard vehicle. It lets the issuer sell many series from one trust backed by a single, changing pool, with each series drawing on the whole pool’s cash flows and income allocated pro rata to the amount each has invested.

A typical card master trust has two notable features. Excess spread is large, because card debt yields far more than the issuer’s funding cost, and excess spread released from the trust can be shared with other series that are short of interest. A financial guaranty is also often added, since recoveries on unsecured card debt are low and there is no collateral to fall back on.

The three credit card metrics

The delinquency ratio is the value of card receivables more than 90 days overdue as a percentage of total card receivables. It gives early warning of deteriorating portfolio quality. The default ratio takes the card balances charged off over a period and divides them by the total pool of card receivables as it stands when the period closes. Read side by side, the two ratios gauge credit losses on the pool, and they are usually linked to early amortisation triggers, so they must be reported throughout the deal’s life.

The monthly payment rate (MPR) is the share of the pool’s principal and interest repaid in a period, measured as collections over the outstanding pool balance. Because card balances are non-amortising, MPR is not a prepayment measure; it shows how fast cash is coming back. The rating agencies require every non-amortising ABS to set a minimum MPR as an early amortisation trigger: if payments slow below it, the deal starts repaying investors.

Mortgage pools come in great variety, with different durations and both fixed and floating rates. Agency MBS are usually pass-throughs, where investors simply own a share of the loans’ cash flows. Non-agency MBS, including commercial MBS, are tranched into senior and subordinated classes, and principal losses are absorbed in reverse order of seniority.

Debt service coverage ratio

Commercial MBS differ from residential MBS in one important way: the loan is non-recourse, secured only by the property, so the property’s own income is what repays it. The DSCR is the key credit measure.

A DSCR below 1.0 means the property does not generate enough cash to meet the required debt payments.

Weighted average coupon and weighted average maturity

The weighted average coupon (WAC) is the pool’s average mortgage rate, with each loan’s rate weighted by its outstanding balance. It changes as loans repay, but at any moment, compared with the net coupon owed to investors, it shows whether the pool can pay. Its companion, the weighted average maturity (WAM), is the balance-weighted average of the loans’ remaining terms, in months. A WAM based on stated maturity avoids guessing whether loans will actually run to term, and it captures the liquidity risk of long, more volatile securities; a WAM based on reset dates instead shows how prepayments shorten the effective term.

When principal comes back gradually, quoting a final maturity date says little about how long the investor’s money is actually tied up. The weighted average life weights each period by the fraction of principal still outstanding, so it measures the average time a dollar of principal stays invested. It lets investors compare an amortising note with other investments of similar maturity, which is why it is a natural measure for MBS and other pass-through structures.

The pool factor is the repayment weighting applied to the outstanding notional. In the table below, each row covers one quarterly interest payment date (IPD) and the period that follows it. The pool factor column is printed to two decimals, while the product column uses the unrounded factor, outstanding balance divided by the opening 89,529,500.00.

The outstanding column shows the balance after each date’s payment; the final 16,470,588.00 is repaid as a lump sum on 24/07/2008.

MBS cash flows cannot be projected without a view on prepayments. The scheduled payments do not change, but borrowers who repay early return principal sooner, and for a given price that lowers the yield on an MBS bought at a premium and changes the timing of every later cash flow. Two common ways to express prepayment speed are the constant prepayment rate and the Public Securities Association benchmark.

Single monthly mortality and the constant prepayment rate

The single monthly mortality (SMM) is the proportion of the pool that prepays in one month: prepayments over the outstanding balance at the start of the month, after the scheduled principal payment. An SMM of 0.65% means roughly 0.65% of the balance remaining at the start of the month, less that month’s scheduled principal, prepays during the month. The constant prepayment rate (CPR) annualises the monthly figure by compounding survival over twelve months:

CPR reflects the characteristics of the pool and the expected economic environment, because it measures prepayment in a given month relative to the outstanding balance.

The PSA benchmark

The Public Securities Association, now merged into the Securities Industry and Financial Markets Association (SIFMA), defined a benchmark that captures how prepayments rise as a pool seasons. A pool runs at 100% PSA if its CPR starts at 0 and climbs by 0.2% a month, reaching 6% at month 30, after which it holds flat at 6%. Other speeds are multiples of this path: 50% PSA tops out at 3% and 150% PSA at 9%. The benchmark assumes new mortgages prepay slowly during their first 30 months of seasoning.

Each metric suits particular collateral. Prepayment measures matter for amortising loans that borrowers can repay early, cash-collection measures matter for revolving card debt, and property income matters for commercial mortgages. The table pulls the measures together, grouped by the asset classes they serve.

After the asset-backed commercial paper market imploded in July and August 2007, investor demand for ABS all but disappeared. Interbank lending dried up to the point where even AA-rated banks struggled to borrow for more than one month, and the strain became acute after Lehman Brothers collapsed in September 2008. The US Federal Reserve and the European Central Bank (ECB) moved first, followed by the Bank of England (BoE), widened the collateral they would accept, making ABS, including MBS, eligible at their daily liquidity windows.

The intention was to let banks borrow against ABS already on their books. Banks quickly went further and created new securitisations of illiquid assets such as mortgages and corporate loans, bought all the notes themselves, and pledged the senior tranche at the central bank’s repo window. These in-house transactions are often called ECB-led deals.

ECB eligibility criteria

An ECB deal is structured like any other, except that it must meet eligibility rules, it faces haircuts, and it can ignore outside investor preferences. The minimum requirements were:

• a public triple-A rating at first issue, and only the senior tranche could be repoed;
• no exposure to synthetic securities: the assets backing the notes could not include, actually or potentially, credit-linked notes or similar claims created by transferring credit risk through credit derivatives;
• a public presale or new issue report from the rating agency;
• a European listing, such as on the Irish Stock Exchange, and book-entry settlement in Europe (for example through Euroclear or Clearstream).

Haircuts and other structuring choices

Euro-denominated CLO securities took a 12% haircut regardless of maturity or coupon. If a note had no trading price in the past five days, or its price had not moved, the ECB applied its own 5% valuation markdown, which equates to an extra 4.4% haircut, and it used its own valuation rather than the borrower’s. US dollar CLO securities took the normal haircuts and on top of them an extra initial margin of between 10% and 20% for currency risk.

Other features were permitted. The deal could include a revolving period, which outside investors would normally dislike. It could be a simple two-tranche structure, with an unrated junior tranche subordinated even to topping up the cash reserve. An off-market interest rate swap was allowed, with penalties if the swap provider was in-house, but any currency swap had to be with an external counterparty. At least one rating was needed, while the BoE required two.

The originator must also decide between a replenishing and a static portfolio. Because these deals are financing transactions, the bank keeps the right to sell or refinance assets whenever better terms appear, usually with no restriction provided the issuer receives at least par, adjusted for accrued interest; the rating agencies accept this. If assets are likely to leave the pool, replenishment keeps the structure efficient by adding new assets instead of running the deal down and building a new one. But rating agencies then examine the bank’s ability to manage the portfolio, and they assign lower recovery rates to a managed portfolio than to a static one, which affects documentation and the bank’s representations.

A hypothetical but realistic case shows the accounting. To secure term liquidity and cheap funding through the ECB facility, XYZ BANK securitises about €2 billion of its corporate loans, as many European banks did in 2008. The deal is entirely in-house: XYZ BANK buys every note, so it is a funding exercise, not a revenue-generating one. Unusually, it includes an extra SPV so that the parent group can reuse the vehicle for later issues, and it has no currency swap, because the notes are issued in three currencies (US dollar, euro and sterling) matching the loans. An external currency swap would have been costly and an in-house one was barred by ECB rules.

Three entities are involved: XYZ BANK itself; a purchasing vehicle, the Master Series Purchase Trust Limited, called the Trust; and an issuing vehicle, the Master Series Limited 1, called the Issuer. At closing, all three steps happen at once and in contemplation of each other: XYZ BANK makes a true legal sale of the loans to the Trust for cash; the Trust issues pass-through certificates that the Issuer buys for cash; and the Issuer issues re-tranched ABS securities that XYZ BANK buys for cash.

Must the bank consolidate the SPVs?

IAS 27 requires consolidation of entities the reporting entity controls, and SIC-12 extends this to SPVs where the substance of the relationship shows control. XYZ BANK must consolidate both vehicles, because in substance their activities run for its business needs, it holds the decision-making power (albeit through an autopilot mechanism), it has rights to most of their benefits and so bears their risks, and it keeps most of the residual risk. That conclusion holds even though third parties, through a charitable trust structure common in SPV arrangements, own the share capital of the Trust and the Issuer.

Does the true sale achieve derecognition?

No. A legal sale is not enough: under IAS 39 an asset is derecognised only if tests of risks and rewards, together with tests of control, show that the economic exposure has passed to someone else. XYZ BANK owns every note, of every rating and level of subordination, with no plan to sell any, so it keeps substantially all the risks and rewards of the loans, and any loan loss reaches it through lower interest or principal on the notes it holds. An in-house deal therefore has no practical accounting or risk transfer effect, and the loans stay on the balance sheet.

Closing-day entries and balance sheets

With the SPVs consolidated under SIC-12 and the sale failing derecognition under IAS 39, the day 1 entries in the individual entities are as follows.

The Issuer’s notes were listed on the Irish Stock Exchange but can still be classed as loans and receivables, because under AG71 of IAS 39 an instrument is quoted in an active market only if prices are readily and regularly available and reflect actual, regular arm’s-length trades. Notes that XYZ BANK holds and never trades do not qualify.

Assuming a 2 billion transaction, on day 1 XYZ BANK shows the notes (investment in the Issuer) of 2bn as an asset and an inter-company borrowing from the Trust of 2bn as a liability. The Trust holds cash X and an inter-company loan to XYZ BANK of 2bn, total assets 2bn + X, funded by pass-through certificates of 2bn and equity X. The Issuer holds cash X and the Trust’s notes of 2bn, funded by its tranched notes (AAA, BBB and non-rated subordinated debt, each XXXm and totalling 2bn) and equity X. On consolidation everything nets out except cash held by the SPVs, if external, and the minority interest in the SPVs’ equity. The loans remain as loans and receivables, and if the notes are later repoed, XYZ BANK shows a third-party borrowing and discloses the collateral.

Swap, transaction costs and capital

An interest rate swap between XYZ BANK and the Trust manages basis risk. Each entity marks it to market, but in the consolidated accounts it nets to zero. One-off costs such as legal and rating fees are deducted from the fair value of the Issuer’s notes and amortised through the effective interest rate over the notes’ life, while running costs are expensed as incurred. In the consolidated accounts these costs were held as an asset, a resource controlled from past events with expected future benefits, justified because the structure secures term liquidity and cheap ECB funding, and tested for impairment at least annually.

If XYZ BANK reports on a solo consolidated basis, intra-group exposures and investments are eliminated, so no further capital adjustment arises beyond the treatment of costs. It must still apply for a waiver from its regulator, the UK Financial Services Authority, under BIPRU 2.1, which would weigh its control over the SPVs, the transferability of capital and assets to it, and the total solo-consolidated amount. The notes it holds carry a 0% risk weighting to avoid double counting, and the SPVs, being unregulated, fall outside the EU Capital Requirements Directive.

A public example: Fast Net Securities 3

Between 2007 and 2009 more than 100 EU banks used in-house securitisations to reach the ECB discount window as interbank funding vanished; Nationwide Building Society was widely believed to have bought an Irish bank in 2008 to gain the euro-area office the window required. Irish Life and Permanent plc originated one public example, Fast Net Securities 3 Limited, which closed in December 2007 and issued two AAA-rated Class A notes and a Class B note, backed by loans to its borrowers and supported by a cash reserve. A AAA ABS note at that time would have been marked at over 200 bps over LIBOR, yet the senior tranche paid only 40 bps over EURIBOR. Because the originator bought every note to use the senior tranche at the ECB, the coupon could be set at a nominal level.

After strong growth from 2002 to 2006, the securitisation market almost stopped in 2007, hit by sub-prime mortgage defaults and the collapse of asset-backed commercial paper. The liquidity squeeze in money markets became a credit crunch and then a worldwide recession. Global, integrated banking and the wide spread of structured credit holdings carried US mortgage losses around the world, and ABS such as collateralised debt obligations (CDOs) magnified the damage from poor-quality US mortgage lending.

The technique still has real merits. It lowers barriers to entry so investors can reach asset markets otherwise closed to them, and by tranching a loan pool into risk classes it lets cash-rich investors help fund large projects. Diversification and customisation widen the pool of buyers and sellers, which adds liquidity and lowers transaction costs for both borrowers and investors. The question is why confidence collapsed.

Off-balance-sheet incentives and the shadow banking system

The same flexibility that made securitisation useful allowed its misuse: banks could move low-quality assets such as sub-prime mortgages off their balance sheets and sell them to investors who barely understood the credit risk. In traditional banking, the bank assesses the borrower, lends and monitors the loan to maturity. Securitisation breaks that link, since the loan is cut up and passed to unknown investors, so the arranger has less reason to be careful. The problem grew when banks built a parallel shadow banking system outside normal regulation. A regulated bank must hold reserves against deposits, but a vehicle funding itself in commercial paper, with no central bank discount window behind it, faces no such rule. Its main risk is that investors stop rolling the paper, leaving it to draw on bank credit lines or sell assets at fire-sale prices, which is what happened in the ABCP crash of August 2007.

Leverage, complexity and ratings

Structured investment vehicles (SIVs) typically ran leverage of about 1:15, and as investors chased yield in a market of tightening credit spreads some reached 1:40 and even 1:50. Long Term Capital Management (LTCM) was at 1:30 when it failed in 1998 and caused serious disruption, so 2007 to 2008 was in effect hundreds of LTCMs failing at once, all more leveraged and all in the same trade. Products added leverage too, as CDO-squared structures invested in other CDOs.

Some products became black boxes that outsiders could not value; their mark-to-market value moved not only with tranche spreads but with changes in correlation inside the portfolio, which affected tranches differently. Rating agency methods looked statistically rigorous but were not widely understood, and with hindsight they were too optimistic, since their models treated a national or international housing correction as almost impossible. Benign conditions and a decade of rising house prices masked the decline in lending standards.

Accounting and liquidity

Liquidity was badly overestimated. Investors assumed AAA securitised paper would trade as easily as plain vanilla AAA paper and could be funded with highly liquid commercial paper, so a huge carry trade of long assets on short funding built up. When sub-prime losses appeared, SPVs had to unwind, and managers discovered a liquidity premium they had ignored. Mark-to-market accounting then created a downward spiral: with no secondary market in a flight to quality, ABS were marked down almost daily, some close to zero, forcing banks to take capital hits unrelated to the actual performance of the loans.

New issuance shrank sharply as a result. In a recession driven by a banking crisis, the flight to quality hits structured products such as ABS before plain vanilla instruments, and a revival depends on rebuilding investor confidence. The benefits remain clear: banks that diversify their funding through securitisation can also diversify their assets, reducing their concentration in residential mortgages and real estate, so the technique stays a valuable way to manage both sides of a bank’s balance sheet.