Friday, 8 June 2012

A really scary story

There was a scary headline in the FT yesterday:


And the report went on to explain:

"The 19 largest US banks are at least $50bn short of meeting new capital requirements under the Basel III accords, according to rules proposed by the Federal Reserve.
The biggest among them would probably need billions of dollars more by the 2019 deadline to comply fully with the rules. Smaller US lenders are about $10bn short of the requirements, the Fed said on Thursday."

Terrifying. And complete rubbish.

These are capital requirements that don't come into force until 2019. The third paragraph of the report says:
"....most banks should be able to reach the new levels by retaining earnings during the next few years rather than by raising capital in the market".
So NORMAL activity will be enough for most banks to meet the new requirements. No shortfall, then. Nothing to look at here. Move along, now.

But in trying to make a scary story out of nothing at all, the writer of this article has completely missed the real issues. There really is a scary story here, but it isn't the risk that banks won't meet the new capital requirements.  That's a red herring.

The real issues are further down the article. Here's scary issue number 1:
"Some banks will be able to calculate their needed levels of capital using internal models, subject to Fed approval"
If you want scary, this should do it nicely. The risk weightings of the complex instruments that failed in the 2008 financial crisis were calculated using banks' own internal models. Using a UK example, both RBS and HBOS were given permission by the FSA to use their own models to calculate risk weightings, and it is fair to say that the FSA hadn't a clue how these models worked. What we do know - now - is that the models allowed the banks to assign far lower risk weightings to the instruments than their actual risk turned out to be. Capital requirements rely on calculation of risk weights - the higher the risk weight, the more capital required to support the asset. So instruments that were actually highly risky were assigned low risk weights by banks using their own models, which allowed banks to reduce their capital - in some cases to dangerously low levels - and the regulators were so impressed with these models that they didn't bother to investigate whether the results they churned out accurately reflected the risk of the assets.

You'd think we would have learned by now that there is IMMENSE moral hazard in allowing banks to use their own models to value and risk weight assets for capital allocation purposes. So why on earth are Fed allowing banks to continue to do this? And the usual defence of this practice - "only the banks really understand these instruments so it's appropriate to use their models" - simply won't wash any more. Regulators MUST understand these instruments, so that they can verify the valuations and risk weightings assigned to them. If the instruments are so complex and so poorly documented that regulators can't understand them, THEY SHOULD NOT BE TRADED.

And here's scary issue number 2.
"Banks will have to hold the same amount of capital against their holdings of sovereign debt issued by Germany as they would for Portugal, Spain and Ireland"
What is happening is that banks are being prevented from using credit rating agencies to provide a measure of risk. If this led to banks doing their own due diligence on investments that would be a good idea.  But that's not what the article is saying. It is actually suggesting that the credit ratings agencies' ratings should be replaced with....nothing. No assessment of default risk at all.  Now, it is not in a bank's interests to tie up capital on investments that are low-risk and therefore give a poor return. If the bank has to allocate the same amount of capital against German debt as it does for Portuguese, but the Portuguese debt pays higher interest, which one will it buy? We have played this scene before. This is how the weaker sovereigns in Euroland got themselves into so much trouble - banks bought their debt in preference to stronger ones because it carried the same risk weighting but paid them a bit more. If the reporter has got this right, it suggests that the Fed (and maybe even the Basel Committee) have learned nothing from the sovereign debt crisis, just as issue no. 1 suggests that they have learned little from the 2008 investment banking collapse.

And finally - scary issue number 3. For this we have to return to the start of the article.

"....most banks should be able to reach the new levels by retaining earnings during the next few years rather than by raising capital in the market".
Now, I know I said this wasn't scary. But it is - just not in the way that the article suggests. You see, if banks are retaining earnings in order to shore up their capital buffers to meet higher regulatory requirements, they WON'T LEND - especially not to riskier prospects. Lending requires capital: every loan on the banks' balance sheets ties up a certain amount of capital, and the riskier the loan the more capital it ties up. In America prior to 2008, banks freed up capital by selling loans for securitisation. But the mortgage-backed securities market is moribund and other securities markets are much smaller. So banks are already having to keep more on their balance sheets than they did before, which ties up capital and restricts lending. If they have to retain earnings to build up their capital, they will lend even less. The US should beware. Raising capital requirements when the securities markets are not functioning properly and shareholders' returns are down is asking for a credit crunch. Banks have seven years to meet these requirements. That means seven years of restricted lending. Does the Fed have any plans to offset the economic impact of what will be a long-drawn-out credit crunch?

Putting all this together we can see that there is a scary story here which the FT writer has completely missed. Banks will continue to be allowed to define their own risk weightings, inevitably leading to insufficient capital allocation against risky instruments - just as in the run-up to the 2008 financial crisis. Banks will be encouraged, through one-size-fits-all capital requirements, to buy the sovereign debt of weaker nations in preference to stronger ones - just as in the run-up to the current sovereign debt crisis. And in order to build up capital buffers, banks will be discouraged from lending to poorer risks such as SME's - just when the economy really could do with an expansion of SMEs to create jobs. Wonderful.

Mind you, if you think the US story is scary, just look at this chart:

(h/t cigolo and @nr_zero for this one!)
These are, of course, European banks. And their capital shortfall is MUCH worse. The total shortfall is 77bn Euros just for these banks, which is by no means all of them: when smaller banks are included, the total shortfall is a massive 178bn Euros. Deutsche Bank alone is short of 10.5bn Euros. And no-one (except, to be fair, the UK) has any serious plans to recapitalise them.

That's the REALLY scary story.


  1. Re Issue 1: RT @Ian_Fraser: Banks internal reporting so weak they're flying blind, warns BoE's Haldane

  2. The problem with capital requirements is how you calculate them, I don't think any method really works, or isn't arb'able.

    Whether you have fixed weights by asset class, use firm or regulator model, risk will build up where its relative requirement is the least.

    Personally I can't see the FSA using a model that would be any better at the peak of the next lending boom, nor can I see firms doing a much better job, and fixed ratios can be even worse.

    Not that I am pretending to have a solution.

    1. Isn't that partly what makes this a scary story?

    2. Yes and no, maybe worry is too strong a word for me. I do agree with your analysis of the article.

      I'm not particularly worried about financial systemic risk at the moment, particularly in the USA.

      I think the solution to a lot of these capital issues reside in accepting whatever the regulatory structure there will be failures, and trying to make those failures happen sooner and in a less painful way. ie accept some ships will sink so build better life boats.

      The biggest taxpayer losses in the USA has been on AIG, an insurer that had more oversight than the banks. That and the auto companies.

    3. Umm, because of the lovely free market in regulators they have in the US, AIG was able to choose a regulator that didn't ask awkward questions about its derivatives trade - mainly because it didn't understand them....So much for oversight. I'm really not at all sure why the US government bailed out car manufacturers though. I suppose they were worried about Detroit. Who'd blame them?

      Actually I think the biggest losses were at Fannie and Freddie, which also had the most watertight Federal guarantees. I'm sure there's a lesson in that somewhere.

      I absolutely agree with you about needing a better resolution regime. Rather than trying to prevent things going wrong, we develop better ways of managing them when they do. Far more sensible.

      The final chart is the point, though. European banks are in a much worse mess, and they are in the eye of the storm. And there is an incestuous relationship between banks and sovereigns in Europe which is extraordinarily difficult to unpick. There is a real possibility of catastrophic bank failure in Europe, I think.

    4. You are entirely correct, Fannie and Freddie were the worst, but they're a whole level of insanity on their own.

      Autos got bailed out because of the massive political contributions from the auto unions.

      I know Europe gets a lot of the headlines and worries, as it should, but the thing that really keeps me awake at night is the risk of a deep Chinese recession.

    5. Totally agree with you about the Chinese risk. And India, and Brazil. The powerhouse is running at stall speed. Heaven help us if it stops.

  3. Since bank regulation does not work, how about this for a much simpler solution.

    Where depositors want their bank to lend their money on for commercial purposes (e.g. to businesses, for mortgages, etc) there would be no taxpayer guarantee that depositors get their money back. After all, such depositors are into commerce, and it’s not the taxpayer’s job to subsidise commerce. Plus the promise that banks make to return £100 for every £100 deposited is plain incompatible with lending that money on in a less than 100% way. It's a self contradiction. It's a confidence trick.

    Any such money deposited in a bank would be similar to money put into a unit trust.

    In contrast, where depositors want 100% safety, banks would be prohibited from lending it on, and would simply deposit the money at the Bank of England where it will earn little or no interest, but taxpayers WOULD underwrite those accounts.

    There would be no instant access to the former type of account (reflecting the fact that the money has been locked up in a business or mortgage) whereas instant access WOULD BE allowed to the latter.

    If the bank goes bust, the 100% safe folk would still have their money. As to those who wanted to engage in commerce, they’d lose theirs. But that’s no different to those who invested in Robert Maxwell and lost their money: the economy did not collapse as a result.

    1. Positive Money have suggested this, and I have some sympathy for it. My main issue is selling the idea to ordinary depositors. People simply don't understand the relationship between risk and return - that if they want a return on their money they must put it at risk. It is Small depositors expect to put their money "somewhere safe" AND receive interest on it - and currently that's exactly what happens with deposits under £80,000. I'm not at all sure how you would sell the idea of having to choose between safety or interest. It would look like a very big rip-off. We have massively dropped ourselves in it with deposit insurance, which has broken the link between risk and return.

    2. There already is NS&I. In a boom people will take risk, even if they know that it is not guaranteed and still expect the taxpayer to bail them out, eg. Northern Rock and Icesave.

      Calomiris has written some great pieces on the hazards created by deposit insurance.

      On the flip side though, the Great Depression clearly shows the huge cost to everyone from a liquidation.

      I personally think the only real solution is to hope to recoup some of it with a small tax on all credit.

  4. Frances, I agree that we have “dropped ourselves in it with deposit insurance” as you put it. And I agree there is a big problem in persuading depositors they can’t have their cake and eat it: i.e. enjoy the benefits of acting in a commercial manner while carrying none of the risks.

    However, where one has made a mistake in the past, I think one has to start rectifying the situation at some stage. Maybe the situation can be rectified a bit at a time, rather than doing it in one fell swoop.

  5. Ralph, Black Raven

    I'm favour of charging depositors directly for insurance, actually. Continue to pay interest on deposits but add an optional insurance charge (which would wipe out most of the interest). Make it clear that if the saver opts out of the insurance they will not be protected in the event of bank failure. That way it is their choice: if they want higher returns they accept the risk that they lose their money; if they want safety they accept much lower returns.

    I am also in favour of taxing credit. This could be done either as a transaction charge or as a tax on balance sheet asset expansion. I slightly prefer the latter.

    1. I don't think the choice between transactional or balance sheet is near marginal.

      Taxing transactions reduces liquidity and therefore reduces asset values. There is also going to be a meaningful juristictional leak.

      Whereas a 'balance sheet' charge (I'd prefer it to be liabilities and try to include all off balance sheet exposures as well) actually gets to where I believe most tail risk comes from; leverage.

    2. I'd prefer it to be assets. Taxing liabilities raises funding costs and reduces returns to depositors - it's really a tax on saving.

    3. True, it is a tax on saving, when it is the form of bank deposits/being a creditor.

      Surely the only difference then between an asset tax would be firms with higher equity would be relatively taxed more when they probably are less of a systemic risk? ie a bank that had no leverage would attract the same tax as a bank with the bare minimum of capital?

    4. I don't see the purpose of an asset tax being to control leverage. I see it as being to control credit expansion, which if excessive is damaging to an economy even if capital levels are good. To encourage banks (and companies) to keep good levels of capital we need to remove the preferential tax treatment of corporate debt, not tinker with liability taxes applied only to banks.

    5. I agree with you that credit expansion is the danger, and that liability taxes shouldn't only be for banks. I'm probably missing something, but I still don't see why you'd rather tax the assets and not the liabilities, surely asset growth backed by equity investment isn't as dangerous as another turn of leverage?

    6. Excessive asset growth, even if backed by equity instead of debt, and even if collateralised, is economically dangerous. Therefore we wish to restrict it. Asset taxation would therefore be a form of Pigouvian taxation. I'm in favour of taxing liabilities too, but the best way of doing this would be to change the tax treatment of interest paid on corporate debt.