Kevin Dowd, 5 December 2018
Dean wrote positively yesterday about Jonathan Ford’s Financial Times piece “The Illusion of UK bank capital strength,” published on 2 December 2018.
Jonathan is right to warn about UK banks being under-capitalised and it’s good to see a journalist of his standing picking up on this issue.
It is however disheartening to see some of the disgraceful abuse made in the ‘comments’ section underneath his article. “Please keep comments respectful,” the guidelines say. “By commenting, you agree to abide by our community guidelines and … terms and conditions.”
Maybe the FT should consider introducing a moderation process to filter out such abuse in future.
Turning to the subject at hand:
Let’s first grant that the RWA measure is useless. It has shown itself to be highly gameable and RWA-based metrics gave no warning of the impending Global Financial Crisis (GFC). It has also been subjected to withering criticism, most notably from the Bank of England’s own chief economist, and in various other studies too (e.g., here and here). It follows that the CET1 ratio, the ratio of CET1 capital to RWAs, is unreliable and should be discarded.
We are then talking about a capital ratio with Total Assets (TA) or some similar variable in the denominator, and the obvious choice is the TA itself.
We still however have to choose the capital measure, the numerator in the capital ratio.
Let’s consider possible choices in the context of the data for Barclays Bank shown in the following table:
Table 1: Capital Metrics for Barclays Bank plc
| Variable | Value |
| Market capitalisation [1] | £28,896m |
| Price-to-book ratio [1] | 44% |
| Book value of equity [2] | £65,672m |
| Total assets [3] | £1,171,000m |
Sources: [1] ShareTelegraph, 28 November 2018. [2] Product of market capitalisation and price-to-book ratio. [3] Barclays Bank plc September 2018 Pillar 3 report.
Unbiased valuation
Assume for the moment that we are interested in an unbiased valuation.
The natural starting point for a capital measure would be the Book Value of Equity (or Book Value for short), which is £65,672m. Given that TA = £1,171,000m, then the ratio of BV to TA = £65,672m/£1,171,000m = 5.61%.
The problem however is that this ‘book value ratio’ ignores the information provided by the bank’s low price-to-book (P2B) ratio, which is equal to 44%. The P2B ratio is the ratio of the bank’s market capitalisation to its BV. The P2B ratio being so far below 100% would seem to suggest that the market does not believe the book value. Indeed, the market is discounting the book value by 56% and that is some discount.
A further reason to take the market values seriously is provided by the Efficient Market Hypothesis (EMH) which maintains that market prices summarise all available information including information not contained in book values. The EMH comes in various forms, but even a weak-form EMH suggests that market values are providing some useful information.
So it would seem that there is a good argument that we should use market values instead of book values, but at the very least, it would be imprudent to rely only on book values and completely ignore market values, especially when market values are signalling apparently large problems that the book values are not picking up.
How then do we obtain the market value of equity?
Easy.
Use the market capitalisation (or market cap), which tells us what the market thinks the bank is worth.
Market cap = £28,896m, so the market-value capital-to-TA ratio = market cap/TA = £28,896m/£1,171,000m = 2.47%.
It seems to me that a bank with such a low market-value capital ratio is not as strong as it might be. If the market-cap-to-TA ratio is 2.47%, then a loss of under 2.5% of its TA would be enough to wide out all its capital.
Alternatively, we can say that the bank’s market-value leverage >=1/2.47% = 40.5. Such a degree of leverage strikes me as very high. After all, wasn’t high leverage a major contributing factor to the severity of the GFC?
How high should banks’ capital-to-asset ratios be in order for them to be regarded as capital adequate? In a famous FT letter in 2010, Anat Admati and 19 other distinguished academic finance experts recommended a ratio of at least 15%, and I could name a considerable number of others who would agree with them. And in his book, The End of Alchemy, former BoE governor Mervyn King suggested that a 10% ratio of capital to assets would be “a good start”.
By either of these capital standards the Barclays ratios are very low. The book value capital-to-assets ratio is barely half the lower King standard and a third the higher Admati standard. Using the market value capital ratio, Barclays is less than a quarter of the King standard and a less than a sixth of the Admati one.
The picture is not much different (well, in fact it is worse) if one uses the Tier 1 leverage ratio (the ratio of Tier 1 capital to leverage exposure, the latter being an alternative to the TA measure) that the Bank of England likes to report. The comparative results are shown in Table 2:
Table 2: Book vs. Market Values, Market cap/TA vs. Tier 1 Leverage Ratio
| Book Value | Market Value | |
| Market cap/TA | 5.61% | 2.47% |
| Tier 1 Leverage Ratio | 4.93% | 2.17% |
Sources: ShareTelegraph, Barclays Bank plc September 2018 Pillar 3 report.
Consequently, I don’t see how one can avoid the conclusion that Barclays is (very) undercapitalised.
I don’t wish to single Barclays out, however: one would get broadly the same conclusion for any of the major banks.
From which one can conclude that Jonathan’s FT story is right.
His corollary also follows: if banks are under-capitalised, then they should be building up capital instead of finding ways to distribute it to their shareholders.
In my option, the PRA should be blocking any distributions or share buybacks until banks have built up their capital to much higher levels than they currently are – and yes, I am firmly in the Admati camp here.
Mark Carney says that the long road to higher capital is over. I say it has barely begun. There is a long, long, long way to go.
And if banks capital levels are still too low, then clearly the conclusions (‘banking system fixed’ etc.) that the Bank of England draws from its stress tests must be wrong too. Not just wrong, but demonstrably wrong, because the empirical capital ratios pull the rug from under any such conclusions.
A prudent assessment
I had earlier assumed that we wanted an unbiased and we do often want that, but sometimes we don’t.
There are situations in which we are concerned with prudent assessment, i.e., where we want valuations or projected valuations or risk assessments that are conservative or biased on the prudent side. Examples are where we are trying to assess a bank’s projected solvency in the event of a major crisis, or where we are conducting or interpreting a stress test that postulates a severe stress, e.g., as in the BoE’s (not-so-) ‘doomsday’ stress tests.
Suppose we are interested in a severe stress and are concerned that some capital items will be unreliable, i.e., they will not be fire resistant, in the heat of a crisis. We saw this happen in the GFC. These unreliable items would include goodwill, deferred tax assets and (perhaps a little controversially) Additional Tier 1 items i.e. Contingent Convertible or CoCo bonds. (Their hybrid capital predecessors proved to be very unreliable the last time round so why are we relying on similar instruments again?) We then take our market cap and decompose it as follows:
Market cap = fire-resistant market cap items + non-fire-resistant market cap items
and we want to eliminate the latter items. Now if only we could measure one or other of the terms on the right-hand side …
But we can! The first term on the right-hand side is (more or less) the market value of the CET1 capital measure, and we can obtain that as the P2B ratio times CET1, bearing in mind that CET1 is a book-value item.
All we need to do is replace market cap with the P2B ratio times CET1 capital.
Using the numbers in Table 1, the ratio of market-value CET1 to TA = 44% times £41,744m/£1,171,000m = 1.56%.
But there is an implicit assumption in this calculation: we are assuming that the P2B ratio remains the same over the course of the crisis. Equivalently, we are assuming that the capital items included in the CET1 capital measure have the same ratio of market-to-book values in a crisis as they do now.
It seems to me that such an assumption is unreasonably optimistic, because we would expect bank share prices and hence market valuation to fall sharply in a crisis and any corresponding fall in book values would, at best, be more limited and also lagged, i.e., we would expect the numerator in the P2B ratio to fall more than the denominator. Therefore, we would expect the P2B ratio to fall as well.
We saw exactly that in the last crisis.
Unfortunately, there is no easy way to gauge how much the P2B ratio might fall in a future crisis. This difficulty, however, does not give anyone carte blanche to ignore the prospect of any such fall and one would hope that any good stress test would take that into account. (As far as I can see, the BoE stress tests do not, but it is possible that some such projection is buried somewhere in their weird and less-than-transparent dividend discount model, but who can tell? It is, at any rate, not in their stress test reports.) Those doing stress tests would then have to come up with a plausible projection of the impact on the P2B ratio and I can only wish them good luck. For those of us who are not doing stress tests, we can either make our own guesstimate (good luck us) or just to keep in mind that any results based on an unchanged P2B ratio in a crisis will be biased on the optimistic side, i.e., will be the opposite of prudent.
In which case, we would conclude that our best prudent assessment of the market CET1/TA ratio is that it is less than 1.57% and possibly a lot less than 1.57%.
The results from our prudent assessment are summarised in Table 3:
Table 3: Market-Values CET1-to-Total Assets Ratio
| Assumed Stressed Price-to-Book Ratio | Market-Value CET1/TA Ratio |
| Current P2B ratio | 1.57% |
| < Current P2B ratio | < 1.57% (or << 1.57%?) |
Sources: ShareTelegraph, Barclays Bank plc September 2018 Pillar 3 report.
So depending on what assumptions you wish to make about the stressed P2B ratio, you can get a market-value leverage that is 1/1.57% = 64, or greater than 64 or much greater than 64.
Take your pick.