Vikram Pandit’s address at a Bretton Woods Committee meeting last Friday was long on calls for transparency.
Shadow banking needs to be inside the regulatory umbrella, he said. High capital requirements push lending outside regulatory scrutiny, and regulators need “a level of clairvoyance”.
Solutions, according to Citi’s chief executive, include tougher regulation, particularly for consumers and consumer lending, and the reform of market structures.
“The more information that we can make available to a larger number of participants in the system, the more safely and efficiently the financial system will operate,” he said.
When it comes to Basel III, Pandit believes, rightly, that it is hard to tell whether two banks with identical tier one ratios are equally risky. Banks allowed to use the Advanced approach end up modelling risk in a spectacularly complex manner. Andrew Haldane of the Bank of England estimated that the Basel II risk requirements (which will be largely unchanged under Basel III) give rise to over 200,000 risk buckets.
The FSA ran an exercise in 2009 to see how important this was. The regulator produced a hypothetical portfolio featuring corporate, bank and sovereign exposures, which it asked UK banks to rate using their internal models. The capital requirements held against sovereign exposures differed by a factor of up to 280%.
Pandit’s regulatory solutions also rely on a hypothetical portfolio test to improve risk weighting.
“Right now, loan-loss reserves, value-at-risk, stress-test results and risk-weighted assets are run only against an institution’s actual portfolio, whose composition is known only to insiders and select regulators,” he said. “The results of the tests, therefore, have no common frame of reference.”
He added that knowing how a company’s risk measurements perform against the benchmark portfolio tells the world how its management thinks about risk, and therefore just how conservative or risky its own portfolio probably is.
Perhaps. Actually, comparing the hypothetical risk weighting and the real risk weighting will tell us a combination of two elements.
First, it will tell us how different the real portfolio and the hypothetical portfolio are.
Second, it will tell us how different the risk assumptions are.
What it won’t tell us is how these are related — how much can be explained by different portfolios and how much can be explained by different assumptions.
It might be possible to reverse-engineer some risk models, and it will certainly provide FIG research teams with a fertile job creation scheme as the financial world digests the new ratio and its consequence. But it won’t strengthen transparency while the importance of assumptions remains a guessing game.
As Pandit pointed out, for reasons ranging from complexity to competitiveness, financial institutions are not prepared to disclose their entire balance sheets to the world.
Fair enough. But they could do a lot better without compromising their business interests. Standard formats for results with standard categorisations would help, and so would more granular information across virtually all categories.
EuroWeek doesn’t have any Citi equity, to pick on Pandit’s bank purely at random. But if it did, having some of the following information in the results might be nice:
· Composition of “Other revenue” – down 85%, but what does it mean?
· Anything about the term structure of its liabilities other than a distinction between “long term” and “short term”
· Any geographical breakdown other than by continent. As has become abundantly clear in recent months, Greek exposure should be treated differently to German exposure
· Pretty well any detail beyond the laughably general about the asset side of the balance sheet.
This isn’t to pick on Citi, particularly. It’s not like this disclosure level is rare, and some of the information is no doubt available on application to investor relations. But a little more effort here would make a lot more difference than asking banks to rate portfolios they don’t even own.