Tuesday, 19 October 2010

Does Debt Due Diligence Scale Well Enough?

I've often made the point that we will only curb excessive fees and risk-taking in the financial markets by vastly simplifying products and making them more broadly accessible. In debt terms, think of this as the 'flat' distribution of risk - or parcelling each loan amount into tiny loans at inception, like at Zopa - rather than the hierarchical or vertical distribution of risk in today's bond markets - where a series of separate loans is packaged into 'tranches' that a bank and credit rating agency estimate will perform similarly, and bonds are issued (and derivatives concocted) according to the varying grades of likely default risk.

The critical potential downside to hierarchical risk distribution is being illustrated by the ongoing 'fraudclosure' and 'forced repurchase' problems in the US mortgage-backed securities market. A root cause may be that due diligence on the scale at which loans must be packaged to fuel the existing bond markets may not scale well enough to provide adequate risk estimates, particularly when the loans have a short history or there is a lengthy chain of loan ownership or loan servicing obligations. Of course, we have a similar challenge on a grander scale in the market for credit default swaps and collateralised debt obligations. But that's a layer above where the current problems are occuring.

The 'fraudclosure' problem arises from allegations in a large number of cases that subsequent loan-owners have not satisfied the formal requirements involved in foreclosing on problem loans. JP Morgan analysts reportedly believe that delays in foreclosures while the technical issues are addressed "will damage senior-ranked non-agency mortgage securities, costing as much as 4 cents on the dollar for certain bonds if postponements take six months."

The 'forced repurchase' saga centres on allegations that certain loans that were sold did not meet underwriting or appraisal standards under the relevant debt sale agreements. The same JPMorgan analysts reportedly believe the banks' losses from repurchases of such loans "will likely total $55 billion to $120 billion, or potentially $10 billion to $25 billion for the next five years."

The fact that these risks have gone unnoticed on this scale until recently suggests a substantial flaw in the due diligence methodology employed in the securitisation and/or subsequent 'collateralisation' process. And one wonders whether any different methodology has since been used by those reporting on whether or not they have an issue amongst their existing holdings, and the scale of any such issue. The explanation of one methodology allegedly used by an anonymous 'whistleblower' to package loans into bonds, was published today by Zero Hedge, and makes interesting reading. In short, the person says:

"...we worked with underwriters of the deal to perform due diligence. That is where this process breaks down. They use sampling to verify the makeup of the pools. There is a lot of pressure to get the deals done in a timely manner so they don’t have time to check every asset. The most I’ve ever checked on a deal is 30%. We’ve done some pools that came back very different from what the trader originally told us.


Don’t get me wrong, I’m not saying that all deals are incorrect, most aren’t. I’m saying that many are, and we have no way of knowing which deals are tainted. Fortunately, most deals have been seasoned a bit which make them easier to value, but the foreclosure documentation is just one instance where my shady scepticism has been vindicated. I knew there was shit floating around in the pools we were putting together, but the sampling technique and level of due diligence was never going to clean it out."

In other words, the scale of bond issuance, pressure of time and the cost of 'full' due diligence seem to encourage costly short-cuts which generate hugely uncertain outcomes.

But this seems to be far less of a problem where loans are appraised and parcelled out at inception according to transparent underwriting standards. Lenders' experience at Zopa is instructive here. But I would say that, wouldn't I? Funding Circle and the Receivables Exchange are examples in the small business finance space, as is the idea of reverse invoice discounting.

1 comment:

Thomas Barker said...

I've always through that mortgages would be better sliced up individually, rather than in pools.

If you're buying a stake in an individual thing, you can "kick the tires" as a buyer. Rather than relying on the seller's checking process.

With a mortgage bond you're buying the best judgement of the packager. It's not too different to leaving the mortgages on the bank's books and buying the bank's own bonds.

Related Posts with Thumbnails