When banks securitize loans to each other

By Felix Salmon
September 15, 2009
Richard Barley has a very depressing column today which begins like this:

Bankers and policy makers agree on the need to revive the securitization market to enable banks to roll over a big chunk of their existing wholesale funding.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Richard Barley has a very depressing column today which begins like this:

Bankers and policy makers agree on the need to revive the securitization market to enable banks to roll over a big chunk of their existing wholesale funding.

If that doesn’t make a lot of sense to you, well, it didn’t make much sense to me either. But let’s rewind a little here, and remind ourselves of the two different types of securitization.

The first type, which is the type I think of when I listen to people talking about “the need to revive the securitization market”, is the securitization of existing real-world cashflows, like car payments or mortgage payments or even David Bowie royalties. Bankers get involved in this market primarily as intermediaries: they talk to the owner of the cashflows (the auto finance company, say, or David Bowie), structure a deal, get it rated, and then sell the bonds to investors. For this they get fee income — but in no sense can that fee income be considered “wholesale funding”.

Then there’s the second type, in which banks securitize their own assets — the loans which they made to their own customers. This kind of securitization makes much less sense. During the boom, a lot of this type of securitization took the form of “synthetic CDOs”, where the banks kept the loans on their own balance sheet and did clever things with credit default swaps to seemingly get rid of all the associated risk. We know how that turned out.

Alternatively, banks sometimes decided to get into the “originate to distribute” business, which was pretty much the single biggest cause of the subprime crisis: because they weren’t going to be holding on to most of the loan, they didn’t seem to care much about old-fashioned things like underwriting quality.

Sometimes, banks did just decide to sell off bits and pieces of their balance sheet by selling loans. This didn’t happen very often during the crisis, since no one was particularly worried about balance sheets getting too big. But when it did happen, the banks sometimes decided to turn to the securitization market rather than simply selling the loans to other banks as they had done for centuries. Was that because they wanted to tap bond investors and thereby increase the potential investor base for the loans? Not really. Barley explains the real reason:

U.S. banks have been subject to a leverage ratio all along, making them keen issuers of securitizations to offload assets, but reluctant holders of triple-A-rated assets, which used up balance sheet for little return. European banks, on the other hand, faced no gross leverage constraints but were regulated according to the Basel II capital rules, which encouraged them to load up with triple-A-rated assets because these carried a very low risk-weighting.

In other words, it was all regulatory arbitrage. Securitization didn’t decrease the total amount of leverage in the banking system as a whole — in fact, it probably increased leverage, by parking assets at European banks who could hold them more or less cost-free as a result of those assets’ triple-A credit ratings.

But in any case, running down the list, it’s still pretty hard to see how securitization could really be considered a funding source. The point of securitization is that you sell assets (after converting them to bonds) — you don’t borrow against them. Conceivably synthetic CDOs might be considered a funding source, but surely no one wants to go back to them.

The one thing which is absolutely clear is that banks should never be in the business of securitizing loans and selling them to other banks. If you want to sell loans to other banks, just do so directly, by syndicating or participating out the loan. Going the securitization route is expensive enough that the only way it makes sense is if there’s some kind of regulatory arbitrage going on, and we don’t want that.

Looking at what’s happening to bond spreads, it’s conceivable that there’s now a bid out there again, among real-money bond investors, for securitized loans. If that’s true, then maybe a few banks could start testing the waters. But insofar as the business of banks securitizing loans to each other has now died, it should never be resuscitated.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

AAA rated tranches of loan securitisations can be pledged with the ECB or BOE for funding purposes whereas vanilla loans cannot. Hence the claim that they generate liquidity.

Posted by kaka | Report as abusive

In repsonse to kaka’s comment – the process of securitising and pledging to the ECB was rife last year.

Felix – have you thought about looking at the impact of this? Clearly the banks “reduce their funding costs”, but at who’s expense? (Citizens of the Eurozone I guess.) And what’s the plan to wean them off?

Posted by vk9141 | Report as abusive

Securitization can be a source of permanent financing. Imagine a REIT specializing in investments of MBS (and perhaps the creation and distribution of MBS). Before the great meltdown in 2007-08, this REIT could go out and buy mortgages from generators and banks. These purchases would be financed in the short-term through repo agreements and “warehouse facilities”. I would then take the mortgages, securitize them as MBS and sell off the MBS. I might repeat the process again and sell off the lower tranches of MBS and CDOs. Haven’t I now just financed my purchase of mortgage loans through securitization?

Posted by Brown Ram | Report as abusive

Bartley’s column is behind a firewall, but I’ll just mention that maybe different banks have different methods of wholesale funding? Bill Cohan has said that Bear Sterns was funding itself largely in overnight arrangements, rolling up to $75 billion every night. Not sure of the details but wouldn’t be surprised if these were repos on super-senior securitized mortgage debt.

Posted by Alan Murdock | Report as abusive

You can’t really make distinctions between “good” and “bad” users of securitzation unless you outline exactly which activities big banks shouldn’t be involved in. You’ve been pretty consistent in worrying about banks becoming too large and unwieldy, as well as certain banks becoming too unwieldy. Banks shouldn’t be music publishing companies? OK. Investment banks shouldn’t be making student loans? OK, does Bank of America count? Citibank?

I think that some kind of resurrected Glass-Steagall, which prohibits large, interconnected players in capital markets from making loans, and letting those that can make loans securitize might be a better solution that trying to make distinctions between types of securitization.

Synthetic CDOs are not a funding source (except in relatively rare circumstances). They’re (ostensibly) risk transfer instruments.

Anyway, the point of Barley’s article is that regulations in Europe made it very efficient (capital wise) for banks to invest in securitisations compared to, say, corporate bonds. That’s one reason why securitisations were underpriced relative to the risk (the huge money marekt appetite for SIV paper also played a big part). If a large non-bank investor base develops for securitisation in Europe, then it will still make sense for funding, though obviously it will play a much smaller role than in the past. The problem is that the non-bank investor universe is far smaller and far more risk averse in Europe than in the US.
Under Basel I, there was a significant regulatory arbitrage for issuers as well, as the regulatory capital for a book of loans was often much higher than the economic capital. Securitisation allowed issuers to reduce their regulatory capital requirement to the economic capital. That particular arbitrage has largely been removed with Basel II, as the regulatory capital is much closer to the economic capital for a given type of asset.

Also, in Europe, and in the US after the new accounting rules come in, ordinary securitisation is in fact accounted for as a secured loan. You derecognise the securitised assets to the extent you have transferred risk, but they remain consolidated on the balance sheet.

Why securitise for funding post-crisis? First, to match liabilities to your assets. Most bank funding is relatively short term, but mortgages can last 30 years. Securitisation allows you to fund those assets to maturity, in theory, although in practice most securitisations are called before then and refinanced. And there are only so many covered bonds you’re allowed to issue. Second, to be able to reach triple-A investors. Third, because the whole loan market, especially for prime loans, is not very liquid at all. It would be more or less impossible for a large bank to fund its prime mortgage lending solely by selling the loans, even if it wanted to, and most don’t. They’d like to keep some of the upside, which securitisation allows them to do.

Posted by Ginger Yellow | Report as abusive

I’d add, this Basel-induced hunger for AAA assets in Europe was one reason why the monoline bond insurers enjoyed such explosive growth the last few years.

hi,actully,i get my loan from mr.larry loan agency ….a friend of my introduce me to this loan agency and it help me for more:VIA;E-mail:loan_dept009@yahoo.com