Measuring bank capitalization and leverage:

Clear as mud

Measuring Bank Capital

Bank regulators worldwide are debating various standards for capital and leverage. In reality many of these measures are difficult to apply, because banks can sidestep certain definitions. In addition, the banks continuously lobby to influence the definitions of acceptable capital and thresholds. There is no single reliable metric and no consistent metric.

Bank lobbyists push for risk-weighted capital measures because “risk” is hard to define and allows more flexibility. Who knows what risk weighting will apply to what asset! The Basel framework tries to define such things and keeps growing more complex and muddy. Basel I was 30 pages long; II was 347 pages long. Just after the financial crisis we got Basel III with 616 pages!

The FDIC's Thomas Hoenig endorses a simpler and more transparent measure. Instead of Basel's flawed risk-weighted measures, he prefers tangible equity to tangible assets. His version of tangible equity is a stricter (and I think more meaningful) capital number. Tier 1 numbers, which are what banks and regulators use, are considerably more generous.

It's hard to even find a consensus on the definition of a simple measure: “leverage ratio”. Because banks hide many assets/liabilities off the balance sheet, there has to be a rule to force them to recognize off balance sheet items. How do you do this when banks don't even know the value of their off balance sheet items? In my opinion, the off balance sheet amounts (derivatives) are the biggest uncertainty in these calculations and result in under-stated leverage. Off balance sheet derivatives result in hidden leverage. This is mostly applicable to the very large banks.

Obviously this confusion isn't an accident. Loose capitalization rules, and under-stated leverage are very beneficial to banks. They can pursue aggressive strategies and earn more profit, while under-stating risks to their shareholders and depositors. Even more importantly, because the leverage is under-stated, the country harbouring the bank (such as Canada) is generally unaware of the magnitude of the risk. If the country and its citizens knew the degree of leverage and magnitude of losses they would someday be forced to absorb, they would never allow the bank to exist in its present form.

Those who understand these regulatory metrics, such as former TARP inspector Neil Barofsky, know that bank regulations are not nearly strict enough. Even after the 2008 financial crisis, we still don't have appropriate limits on bank leverage and this is a global problem.

Tier 1 Capital, and the Leverage Ratio (Basel III)

Pull up any large bank's financial statements and you'll probably find at least three different numbers for Tier 1 capital, under different Basel frameworks and transitions. That being said, Tier 1 capital is still a useful number.

The Common Equity Tier 1 capital (CET1) is a new, stricter definition and is what I'll use in my bank risk calculations. This is common shares plus retained earnings, less goodwill & other intangibles, and is the closest standardized number to Hoenig's suggested Tangible Equity number. Canadian and European banks now publish this number, but American banks do not.

Basel III introduces a Leverage Ratio (Tier 1 capital divided by Total exposure), with off balance sheet amounts included. This would have been a good way to assess a bank's overall risk. Unfortunately, it's difficult to calculate given the complexity and uncertainty of derivative exposures (and still involves questionable risk adjustments). In the end, Basel III introduced the Leverage Ratio as a non-binding supplement and will calibrate it until 2018. So it basically means nothing: the off balance sheet part of the exposure is loosely defined, and there is no requirement to obey a limit or to even report a number!

So what can we do, if banks aren't publishing their Leverage Ratios?

The ratios that the banks do publish are risk-weighted numbers. As I described above, these are preferred by the bankers' groups and give the banks more flexibility. I think the non-risk-based leverage ratio is superior because it's simpler and doesn't let a bank play with arbitrary risk categories. There is also evidence that the non-weighted leverage ratio better predicts failures of large banks (see Bank of England speech at bottom).

In the absence of published leverage ratios, I will calculate a variant. I will calculate Leverage = Assets divided by Common Equity Tier 1 capital.

Some important notes:

Leverage formula

My formula isn't perfect, but I think it's more meaningful than the risk-weighted Basel ratios. This formula gets close to Hoenig's recommendation and the Leverage Ratio that won't be implemented for many years. Note that off-balance sheet exposure is not included.

Leverage =


Common Equity Tier 1 capital

Links and references:

- Perpetual Bull,