Breakingviews on Twitter
Search League Tables

Thursday, 30 October 2014

Model behaviour

How regulators can restore trust in bank capital

The Basel Committee just dealt another blow to the self-assessment of banking risk. This powerful group of global supervisors published new evidence on July 5 unmasking shortcomings in the models that tell lenders how much capital they need. For fans of the status quo, it makes gory reading.

Regulators currently allow banks to weight their assets by how risky they are. This then determines how much capital they need. But investors increasingly believe that lenders massage their sums to cut their capital requirements. From the outside it can be difficult to see where justified diversity of risk ends and where fiddling begins.

The committee has made a stab at cutting through this. It looked at 40 percent of the banking book assets of 32 banks, assumed they each held core Tier 1 ratios of 10 percent on their risk-weighted assets (RWAs), and tested how they would fare if each had to assess their assets using the average risk-weightings of all the banks in the study. One unnamed European bank winds up with a 7.8 percent core Tier 1 capital ratio, while an American peer rises to 11.8 percent. Apply the same assumptions to the whole banking book and the worst bank falls to 5.9 percent, with the best at 15.7 percent.

Not all of these differences reflect foul play. The appropriate risk weight for a Spanish mortgage will differ from a similar asset in Germany. What investors are rightly worried about is that the “own model” system allows banks to assign different risk weights to exactly the same loan – and that their own national regulators are complicit.

Basel’s study is timid on policy prescription, but it strengthens the case for standardisation. The question is how to get there. Global regulators could force banks to work out risk weights by means of standardized inputs dictated by them, instead of selected by banks themselves. This could be quite painful: some of the RWAs yielded by banks’ own models in Basel’s study are only 25 percent of their standardized equivalents, which means capital needs would soar.

The committee knows this. The latest thinking is to force banks to disclose both their own model RWAs, but also their standardized figures, according to a person familiar with the situation. That stops short of holding banks to a higher capital standard overnight. But at least investors could then exercise proper scrutiny – and be better placed to ask if its management is pulling a fast one.

Have your say

To have your say, you have to be signed in

Context News

Internal models used by banks to calculate their risk-weighted assets show disparities worth up to 4 percentage points of core Tier 1 capital on the same types of assets, according to research published on July 5 by the Basel Committee on Banking Supervision.

The Committee looked at a sample of 32 global banks and required them to risk-weight 40 percent of their banking books using the average of all their risk weights, and also assume a 10 percent starting core Tier 1 ratio. Although 22 of the banks saw their capital ratios only rise or fall by 1 percentage point, the lowest ranking bank’s capital ratio fell to 7.8 percent, while the highest ranking lender’s ratio rose to 11.8 percent.

The Committee’s report also found that if the same assumptions were applied to all the banking book, the weakest bank’s capital would fall to 5.9 percent, while the strongest would rise to 15.7 percent.

(Launches in a new window)