The possibility of a European minimum over-collateralisation level for all covered bonds has been discussed on many levels.
However the possibility of asset-class specific o/c raises some totally new problems. These need to be addressed before the idea can work.
I have discussed the idea that a minimum level of over-collateralisation for all covered bonds might be enshrined in the new covered bond directive from a couple of perspectives already – whether it makes sense being the most obvious one. But a rumour going around the market that the directive next March might set a minimum over-collateralisation on the basis of the asset class – one level for residential mortgages, another for commercial, a third for public sector assets – is a new angle.
On one level it sort of makes sense – different asset classes have different risk characteristics so it is only reasonable that this should be reflected in different minimum over-collateralisation levels. Also, it is politically pragmatic – a greater alignment between actual risk and statutory over-collateralisation sates some of the complaints against the principle of one size fits all o/c.
But there are a whole load of problems with the proposal too.
Lets take that last point first. Commercial mortgages might be riskier in any given country than residential mortgages. But if you say that the former needs 5% collateralisation when the latter only needs 3% (I’m making these numbers up. I have no insight into the real values) and you leave all of the other glaring collateralisation/risk misalignments un-touched, then aren’t you implying that residential mortgages in country A – let’s say Greece - are now officially safer than commercial mortgages in country B? – let’s say Germany, to be provocative.
Then there is the topic of the relative riskiness of the asset classes being different in each member state. The relationship between residential and commercial mortgage riskiness I can probably accept is fairly consistent across Europe. But in some countries residential mortgages are undoubtedly safer than public sector credits – those countries where covered bonds trade through the government credit curve for example – whereas received wisdom in, again let’s say Germany, is that the reverse is true. So, which country is going to determine the relative minimum o/c of residential mortgages and public sector assets? Greece or Germany?
Then, on top of all of the perfectly justified points that have been made about the calculation methodology differing by state, the definition of commercial and residential mortgages also differs by state. How do you treat mixed-use properties? Or buy-to-let properties? Or housing co-operatives? There are EU wide definitions of what a residential mortgage exposure is, but if these don’t map too closely to the definitions in the issuer’s own IT system (trust me, they don’t) then in the past hasn’t mattered too much in the past.
Finally - I’ll stop ranting soon – if we have a different over-collateralisation for the purposes of the prudential treatment of the bonds shouldn’t we also have a differentiated level for the purposes of the treatment of the associated derivatives under EMIR? Currently the clearing exemption for covered bond related swaps is based on 2% minimum over-collateralisation of the underlying programme no matter the asset class, no matter the member state. Will this also need to be changed?
I hate to appear negative about practical proposals but there are a lot of objections that need to be addressed before the rumoured proposal can be widely accepted.
By Richard Kemmish