In a previous post, we detailed trades that while making no sense economically, allow banks to game regulations around capital requirements and kick the recognition of losses further down the road.
This may have left you wondering how this is possible under the Basel II regulations, so we thought we’d walk you through the relevant sections, and outline the cases where this trade does, and doesn’t, work.
As a reminder, one of the main trades involves pooling together a bunch of assets that are performing poorly (e.g. ABS, RMBS, loans, etc.), then buying protection on the first loss piece — that you expect to lose 100 per cent of its value. The counterparty that sells the protection is willing to participate because their total payment will also be equal to 100 per cent of the loss. It’s pure window dressing of the accounts and arbitrage of the regs.
First up, here’s the table on the securitisation risk weights under the standard approach:
What ‘deduction’ means (cause we’re concerning ourselves with the stinkiest of exposures here, hence we’re at the far right of the table above):
561. When a bank is required to deduct a securitisation exposure from regulatory capital, the deduction must be taken 50% from Tier 1 and 50% from Tier 2…
That is, the entire exposure is subtracted from capital (split over tier 1 and tier 2). But if you were to buy credit protection…
196. The protected portion is assigned the risk weight of the protection provider.
No more deduction!! Instead, there’ll be something like a 20 per cent risk weight applied. That is, from having to hold $100 of capital for a $100 exposure, you’ll hold more like $1.60. Bargain. Nevermind the pesky fact that the pricing of the transaction means that you expect the value of the exposure to go to zero. (This is why the trades with monolines worked so well, the risk weight for facing them was very low since they were rated so highly.)
But, surely all that happened is that the risk of the first loss piece just moved to the bank that sold the protection, right?
Correct:
584. When a bank other than the originator provides credit protection to a securitisation exposure, it must calculate a capital requirement on the covered exposure as if it were an investor in that securitisation. If a bank provides protection to an unrated credit enhancement, it must treat the credit protection provided as if it were directly holding the unrated credit enhancement.
That is, the bank writing the protection is indeed laboured with the same burdensome regulatory capital requirements that the buyer of protection had. So, they should only be willing to shoulder the burden if the fee is handsome enough or if…
606. Banks that have received approval to use the IRB [Internal ratings-based] approach for the type of underlying exposures securitised (e.g. for their corporate or retail portfolio) must use the IRB approach for securitisations.
… their regulator-approved models allow them to hold less capital against the same exposure.
In summary, the only way such deals could have been done is if banks were willing to pay over the odds to window-dress and delay losses, or if there are banks out there who’ve gotten regulatory approval for some rather crafty models.
And there’s a third possibility. Banks could go to the buyside for protection. Hedge funds are gloriously unencumbered by Basel regulations (as were the monolines, which is also why that trade was great, but this route is now closed off, thankfully).
The only limitation with hedge funds is that the weight associated with the counterparty risk for facing an unrated fund is going to be significantly higher than facing a bank (or AAA-monoline). That said, anything less than a 100 per cent deduction has upside.
What’s to stop this going on?
On the one hand, the internal compliance functions of a bank should stop this sort of behaviour. Any risk manager worth his or her salt should recognise that there hasn’t been much in the way of genuine risk transfer here.
On the other, this is a pleasant example of regulators stepping up to the plate to try to stop an abusive practice. Armed with the guidance letters of The Fed and the Basel Committee, supervisors should be empowered to tell their charges that this sort of engineering just won’t fly, and it will not be granted regulatory capital relief.
Consider the guidance a shot across the bow: model shopping for better regulatory treatment is not on, and trading just to window-dress is not on. Let us hope such vigilance continues…
Related links:
More Angry Birds, less regulatory arbitrage, please – FT Alphaville
An unusual letter from the Fed, an unwinding of hedges – FT Alphaville (January, 2011)