Just before Christmas, Basel released its by-now-traditional new set of capital rules for securitizations — the third version, following a second iteration at the end of 2013 and a first pass in 2012.
The changes disappointed the industry, again, with tweaks to the details — but no change to the risk weight floor, which could make senior tranches of securitizations more expensive for banks to hold than underlying loan portfolios.
But Basel’s latest proposal for banks to calculate their own regulatory capital — rather than capital to allocate to their assets — could push the pendulum back the other way.
The new rules, published in a consultation just before Christmas, also feature a capital floor, “meant to mitigate model risk and measurement error stemming from internally-modelled approaches”.
In other words, no matter how safe banks assess a portfolio to be, there will be a certain amount of capital that must be allocated to the assets. It is supposed to be a backstop to make sure that whatever clever modelling banks come up with, enough capital stays in the financial system.
Capital floors came in with Basel I, as a transitional measure, but remain in bank rules right through to the present — still as a “temporary” solution.
However, these limits rarely bind on banks. The new proposals, if they set the capital floor at a much higher level, could sharply reduce bank incentives to use internal modelling of portfolios at all.
What, after all, is the point of a long, complex regulatory modelling effort if it does not demonstrate that a portfolio needs less regulatory capital allocated to it?
As Simon Hills of the British Bankers’ Association puts it: “In extremis banks could stop investing in regulatory approved [internal ratings-based approach] modelling capacity and accept the credit risk capital ‘answer’ set by the standardised approach.”
In other words, if there is no capital reduction to be gained from using internal models because the capital floor cuts in at a high level, why use them at all?
And if banks start abandoning internal modelling for regulatory capital en masse, that cuts the cord between economic capital (how risky banks think their loan books are) and regulatory capital (how risky regulators think bank loan books are).
At the moment, the internal ratings-based approach keeps the concepts approximately in line. Hills suggests it is a “somewhat artificial alignment between regulatory capital modelling and economic capital modelling which now is neither fish nor fowl”.
But it does exist, at least in an approximate form. Where it is far enough apart, banks start selling synthetic CDOs — the good kind, like Standard Chartered’s START programme, which genuinely transfer risk.
Push regulatory capital and economic capital further apart, and the transfer of risk outside the banking system will sharply accelerate.
As Hills says: “This would inevitably highlight differences between regulators’ views of the riskiness, and therefore capital intensity of different classes of assets and banks’ own estimates of capital requirements. And where banks think those requirements are too high a natural response might be to sell such exposures off to unregulated entities in the shadow banking market by securitizing them.”
Perhaps this just represents a levelling of the playing field. Securitization as a format took the heat for the crisis first, and hardest, with a raft of regulation introduced in 2009 and 2010 which could have killed it entirely.
Capital floors for securitization tranches are now being matched by (more lenient) capital floors for bank lending, raising the total capital held against all asset portfolios. That ought to strengthen the system — but will be deeply strange for the securitization industry.
If it all comes to pass, securitization bankers could find themselves giving thanks for the Basel committee and their tough stance.