Capital doesn’t matter if valuations are wrong November 15, 2012 at 7:08 am

Bloomberg has an interesting story on the Danish FSA’s pursuit of mis-stated financials at banks:

Denmark’s financial regulator is warning the country’s banks that an understatement of lending risks won’t be tolerated as it embarks on a hunt to catch what it’s dubbed “backdoor” capital dilution.

The Financial Supervisory Authority will review internal rating models that determine how much capital a lender sets aside to ensure banks don’t find a way around stricter standards. While banks may fulfill capital requirements on paper, recent failures suggest risk weights don’t always reflect reality, leaving buffers too small to absorb losses… When Denmark’s housing bubble burst more than four years ago, it revealed widespread capital shortfalls that have since led to the demise of more than a dozen regional lenders. Toender Bank A/S, the most recent insolvency, followed a reported three- fold increase in profit in the first half and a solvency ratio – - a measure of financial strength — of 17.3 percent at the end of June. Yet an inspection last month by the FSA revealed bad loans almost 10 times as big as those reported by the bank, wiping out its equity.

Bloomberg loses a few marks here for not being precise: the capital requirements are fine, but the capital available to meet them is bigger than it should be because the bank has not taken enough provisions and thus has over-stated its earnings. What is interesting is that when the Danish FSA tightened their standards on impairments, impairment charges doubled at one large (and well-regarded) bank, Nordea. Presumably the impact elsewhere was similar. This kind of solid, boring policing of lending is very important, so kudos to the Danish FSA for doing it. One does wonder what Toender’s auditors were doing though…

3 Responses to “Capital doesn’t matter if valuations are wrong”

  1. David, to some extent underprovisioning will be addressed by the requirement to deduct from capital any shortfall of provisions vs expected losses. But if the model forecasting EL is poor, this is a weakened control. On the broader issues of models etc, I do fear some moral hazard attached to a regulator insisting they know best. But maybe that is less bad than an internal model built by a bank which willnever have a risk appetite lower than the regulator?

  2. QM – yeah, as you say that works if you believe the PDs. I think the Danish issue is slightly different though; as I understand it (which isn’t well), what the Danish FSA are doing is imposing tight standards on current impairments. That is, they are stopping extend and pretend. In my experience banks talk a great game here but then sometimes play fast and loose with implementing their own policies – and of course the auditors just nod and say yes.

  3. Really great article on Deus Ex Macchiato