Counter-cyclical capital

I flatter myself that I was one of the first bloggers (although far from the first academic) to comment on the need for anti-cyclical capital rules. Three years later, this is becoming accepted wisdom. People still seem to think that identifying the cycle is difficult. I'm sure it is not, and I identified a number of indicators that could be used to set capital levels in my book. Now the BIS annual report has reviewed several possible indicators: credit spreads, changes in real credit provision, and a composite indicator that combines the credit/GDP ratio and real asset prices. And, rather unsurprisingly, they all work to a reasonable degree.

The loneliness of the long distance regulator

Lord Turner, a decent, thoughtful regulator, told the Treasury select committee yesterday that

there should instead be a "tax on size" by requiring the big banks to set aside more capital when they expanded beyond a certain size.

Quite right too, and nice to see an idea I championed being mentioned in such august circles. The bad news is that

Turner also warned the MPs that the radical changes to regulations needed in the wake of the banking crisis may not take place because of the emergence of green shoots of recovery and "exhaustion".

Regulatory reform is a marathon not a sprint and I share Turner's doubts that we have the stamina to do a good job at it.

Cash up front please

The Telegraph reports an idea of Paul Tucker's: making the banks pay to clean up their own mess:

The banking industry could be told in advance that, if ever there was another crisis, the ultimate cost would come from banks themselves. In the midst of a crisis, that would not be possible. A government would have to pump in new equity. But when the dust had settled and the government had sold its shares, the loss (if any) could be calculated - and then collected from the industry via a levy.

This isn't a bad idea. But there is a better one. Make them pay before the crisis.

There are various ways to do this. One is to take cash from the banks, via a beefed up version of the way the FDIC works. In order to be a financial institution, you need an annually renewable license, and the license should be expensive.

A more intriguing one, though, is to make the banks hand over each year not cash, but one year call options on their stock. The regulator would then hedge these options. The bank's shareholders would only be diluted if the stock went up, sugaring the pill for them, while the hedging process would ensure the regulator made money whether the stock went up or down. Indeed, as the position is long gamma, a big fall would be particularly profitable to the hedging strategy.

What we need now

Less regulation of financial institutions. More leverage. After all, look how well that worked the last time... No? No. So it is rather a surprise that there is even any discussion of Goldman Sachs moving back from being a commercial bank to being an investment bank. The Reuters story is here. The regulatory regime Goldman used to operate under before the change - in the days after the collapse of Lehman - was so flawed that the SEC ended it. Given that it is not possible to be a CSE any more, if Goldman wanted to shed its bank holding company status and the regulation that goes with it, what exactly would they do?

Monoline death watch, day 1000

(The fate of the monolines was sealed by the risk they wrote in 2005-2007, so day 1000 is if anything conservative - it takes insurers a long time to die thanks to their accounting and the pay as you go nature of the claims against them.)

In a move so predictable it hardly raises a yawn, Bank of America, Citigroup, JPMorgan and 15 other large financial institutions filed suit on Wednesday against MBIA, claiming the bond insurer reduced its ability to pay policyholders by splitting its business in two.

It is difficult to think of how a monoline could split without generating lawsuits. If x of capital and y of investments support z of risk, and you split z into to pieces, how do you divide x and y? There is not widely agreed answer, so someone is going to think that the credit quality of the piece they end up with a claim against is lower that it should be - and they will sue. Moreover since there are clearly diversification benefits between risks, even if the split is entirely fair, the capital needed for the two pieces is larger than that for the original whole.

The lesson? Agree collateral upfront, based on your marks.

How much capital does a bank need?

Probably the best answer is the least amount that gives comfort and confidence to debt holders, regulators, and other stake holders. But that is a moving target.

Over the last twenty years people like me have spent a lot of time trying to construct and improve capital models. At its simplest, a capital model uses some measure of risk to deduce how much capital is required for some portfolio. The problem is that many of these measures of risk have proved highly fallible, and thus capital has been systematically understated. Moreover, thanks (1) to leverage and (2) to the fact that losses are a deduction from capital, even small capital mis-estimates can emperil a bank in a crisis like the current one.

These chickens (or in Citigroup and Bank of America's case, flocks of giant mutant turkeys) are coming home to roost. Six of the largest nineteen US banks require more capital, according to the FED: and you can be sure other banks around the world do too.

What you can be sure of in these discussions is that the numbers are essentially arbitrary. No one really knows how much capital a bank needs at any given point, not least because risk based capital models have lost their credibility. Capital is adequate if and only if it keeps the relevant stakeholders happy: and risk based estimates no longer keep people happy. So don't expect the negotiations of the coming weeks necessarily to make sense. It will all be about the deal that can be done when everyone gets around the table.

Update. Matthew Yglesias, via Gary Becker, has a nice characterisation of what we can expect from capital, and indeed regulation in general:

The best you can hope from a regulatory regime is ... [a] fairly crude rule will improve on the outcomes generated by the unfettered market... when we're looking at a regulatory regime that seems to be working okay, and the regulated parties start saying we need tweaks x and y and z and oh there's no danger there we should be very suspicious. We shouldn't count on being to fine-tune our results to perfection

Another failure to call

From the FT:

Mizuho Financial, Japan’s second largest bank, has opted not to repay a $1.5bn junior security at the first opportunity, adding to the list of [4] institutions globally that have not repaid similar debt and potentially aggravating some of its investors.

The instruments are perpetual subordinated bonds with calls and no-steps, paying 8.375%. This will further spook Tier 2 and hybrid Tier 1 investors.

Good bad/bad bank

Willem Buiter has a discussion of how good bank/bad bank separations might work in detail: the mechanics come from Robert Hall and Susan Woodward. I will simplify the argument a little, and discuss the issues.

Consider a bank with:

Assets		Liabilities
Good loans	1000	Deposits	1200
Bad loans	500	Bonds Issued	600
Other assets	380	Shareholder's funds	80

(Let's ignore the off B/S stuff for this post and assume that all of the other assets are good.)

The proposal puts the deposits and good assets in the good bank, and calls the difference between assets and liabilities 'capital'. Thus we have for the good bank:

Assets		Liabilities
Good loans	1000	Deposits	1200
Other assets	380	Shareholder's funds	180

Notice that the good bank is well capitalised under this proposal.

The bad bank owns all of the equity in the good bank. For it we have:

Assets		Liabilities
Bad loans	500	Bonds Issued	600
Equity in good bank	180	Shareholder's funds	80

It is fairly likely that the equity holders in the bad bank will be wiped out over time, which is right and proper. If the good bank makes money and declares a dividend, the bad bank will receive that income as it stands. Meanwhile the debt holders of the bad bank now have a claim on a rather worse quality institution, at least at first sight. This is a proposal with rather little moral hazard.

The issue comes when we consider the bad bank's position. It is not capitally adequate, not least because material holdings in credit institutions (i.e. its shareholding in the good bank) is a deduction from equity. One might argue that it does not need a banking license as it is now in run off, but still, it is so leveraged that its management will have to sell some of the equity in the good bank. Does a forced seller of bank equity (albeit good bank equity) really help financial stability?

Also notice that the bad bank would consolidate the good bank from an accounting perspective. Again, to get deconsolidation it would have to sell at least 50% of the good bank's equity.

The proposal in short makes sense from a moral hazard perspective, and transfers the taxpayer's deposit guarantee to a well capitalised institution. But it does force the bad bank to sell its position in the good bank almost at once, and that is a rather worrying side effect.

Bigger is worse

A correspondent of mine asks, apropos this story:

AIG, whose reach is so vast that the government warns letting it fail would cripple the very world financial system

Why are companies allowed to get this big? The answer is of course that they shouldn't be, if you care about financial stability and moral hazard, at least. What we need is capital requirements which are a strong increasing function of size, so that larger companies are forced to have a lower ROE than smaller ones, and hence so that they have an incentive to split. The current rules of course are the opposite: the ability to use internal models within Basel, and the advanced capital models that big insurers use, have the effect of making capital requirements lower for bigger companies. You don't need a crystal ball to figure out that that is a bad idea.

Real Solvency and Fantasy Solveny

There is a `the banks are insolvent' meme going around at the moment. It's rubbish, but something close to it may be true.

The reason it is rubbish is that solvency means assets > liabilities under the firm's accounting standards. This is clearly true for all the large banks.

What the commentators mean by `the banks are insolvent' really, then, is `under my idea of what accounting standards should be, the banks would be insolvent'. Clearly this is a little different, however rational the particular accounting counterfactual concerned is.

Let's look at some of the choices in the space of accounting methods.

• Pure accrual accounting would value every asset and liability under accrual accounting, with whatever loan loss reserves the firm can get past its auditors being taken. Under this measure pretty much every bank is solvent.
  
• Pure rigourous fair value would use fair value for everything, with prudent valuation adjustments being taken wherever there is uncertainty. Under this measure, many banks would be insolvent.

Most banks definitions of solvency are closer to the first than the second of these at the moment, of course. I suspect that most commentators who say that the banking system is insolvent are implicitly thinking of something like pure rigourous fair value. In any event, there are many, many accounting standards between these two extremes, of which any given bank's choice is one.

Two more things to note.

First, insolvency implies that the bank is not capitally adequate, but capital adequacy is a stronger constraint. It implies solvency* plus capital > capital requirements**. Losses challenge both solvency and capital adequacy as they erode capital, but the capital adequacy test is hit before the solvency one.

Second, solvency or insolvency have nothing to do with liquidity. A bank can be insolvent and perfectly able to fund itself (if that fact is well enough hidden) and highly solvent but unable to fund.

For further reading, see a good if long post by John Hempton here.

*Actually this is not quite true as the regulatory notion of solvency is not quite the same as the accounting one. Regulators apply a few (typically minor, in the big scheme of things) valuation adjustments to GAAP.

**The definition of Capital is much more country specific than that of Capital requirements. The `Basel 2 capital requirement' is close to being the same everywhere (although there are some differences in national implementations). But the definition of capital varies significantly, especially in the treatment of things like deferred tax assets, goodwill, and unrealised gains on held to maturity positions.

Bank capital? Not so much...

Sam Jones has an excellent post on FT Alphaville picking up on a momentous, and well concealed volte face from FSA. The regulator has (finally) embraced procyclicality.

We have continued to give consideration to appropriate long term changes to the bank capital regulatory framework. We believe that it would be preferable for the capital regime to incorporate counter-cyclical measures which lead to banks building up capital buffers in good years which they can draw down during economic downturns.

The well-concealed part picked up by Alphaville is:

we are amending the variable scalar method of converting internal credit risk models from point in time to through the cycle... These changes will significantly reduce the requirement for additional capital resulting from the procyclical effect.

As Alphaville says:

The FSA is authorising a modelling trick

But a particularly expedient one. Welcome to the bottom of the cycle capital regime. Just one thing. What if this is not the bottom?

Unrealised P/L and Tier 4 capital

Mark to market is great. It gives the users of financial statements the best information available about the value of a company. But, as we have seen over the last year or so, it also has the drawback - at least when applied to banks and such like - of encouraging procyclicality. On the way up, mark to market gains, once audited, form retained earnings, and so contribute to capital. This capital can then support more risk. On the way down, losses reduce capital and so inhibit risk taking just when it is vital for the economy that financial institutions to step up to the plate.

So.... let's split the link between unrealised gains and retained earnings. Specifically, I propose splitting the retained earnings component of tier 1 into two pieces. The first, retained realised earnings, would be as before. The second, retained unrealised earnings, would be the sole component of a new class of capital, tier 4. (Tier 1 is equity and highly equity like capital; tier 2 is reserves and certain types of long term sub debt; tier 3 is short term sub debt.)

Unrealised gains and losses would change tier 4. There would be constraints on the total amount of capital that could come from tier 4, just as there is today on the total that can come from tier 2. Exactly what would work needs some research, but my guess is that, say, having a rule like tier 1 must be 8 times bigger than tier 4 would work. On the way up, this would restrict the benefit available from unrealised gains. That buffer would then be available to absorb losses on the way down without restricting risk taking.

The shocking non-call

A deeply scary event for the capital instrument market occurred today. Deutsche did not exercise a call on a lower Tier 2 instrument, the 3.875% sub notes of 2004/2014. From their press release:

Deutsche Bank has decided not to exercise its early redemption option to call the Notes at par because replacement costs would be more expensive than the existing EURIBOR +88 bps step-up coupon. Accordingly, the Notes' early redemption provision at the option of the issuer is not in-the-money.

The entire capital instrument market is based on the principle that banks will call their sub notes at the earliest opportunity - most investors calculate duration on that basis, and few people calculate an option adjusted spread, viewing the option not to call as reputationally too dangerous to exercise. The world has changed: expect the prices of these things to plummet. It will also become significantly more expensive for banks or insurers to issue hybrid capital instruments. This is really really not good.

Hedging procyclicality

It suddenly occurred to me this morning that you can, theoretically, hedge the cost of procyclicality.

Suppose your credit risk capital requirements in good times are EUR 10B, and they rise to EUR 13B in a crisis. This is roughly right for a large bank: a Deutsche bank study showed between 30 and 80% increases in capital for their investment grade corporate book, while the Bank of England found 15-40%.

The cost of equity capital is somewhere around 12% for a large bank. Let's go with 10%, assuming that you can meet some of it from cheaper Tier 2 capital. That means you need to make EUR 300M a year to be hedged. Hence you want some position that makes money in a recession. Considering the iTraxx, we see that a 150 b.p. credit spread change is perfectly possible: this has actually gone from 25 to over 200 in the recent events. Unfortunately to make 300M from a 150 b.p. change, you would need to buy a notional of 20B. That's quite a bit of iTraxx.

How about equity indices? Clearly the business cycle is quite long: five years peak to trough would be reasonable. So how about buying an ATM five year put on the S&P? Assuming a 20% fall from the high - it has gone down more recently, but 20% is more typical of an average recession - you would need a notional of EUR 1.5B. That's perfectly possible. But the premium for a five year ATM S&P put using pre-Crunch vols is very roughly 20% of notional, and you need to earn that premium back too. Hmmm. You can do better by buying away from the money, longer dated options and profiting from the move in implied vol too, but the notional will be even larger. Also if the recession lasts longer than a year, you need more puts.

So perhaps practically it is not so easy. But the *idea* of hedging increases in capital requirements is interesting...

The pain in Spain stays mainly on the plain

Or perhaps post-industrial wasteland is more apposite, for those of you who have been to Santander's headquarters in Boadilla. Anyway. The self-proclaimed over-capitalised and conservative bank isn't. Well-capitalised that is. There is a 7.2B EUR rights issue, according to the FT.

There are two readings of this. One is that if even a bank as old fashioned as SAN needs new money, the financial system is going to hell in a handcart and practically everyone needs new capital.

The other (not necessarily exclusive) interpretation is that SAN has some accrual accounted nasties on its balance sheet which, if they were fair value accounted, would already have produced substantial write-downs. SAN can only go to the fountain once because the admission that it needs new capital is tantamount to the one that its loan loss provisions are not sufficiently prudent. And once you say that, investors might never trust you again.

Personally, I am in favour of transparency. If I were CFO of a bank with a substantial banking book under accrual, I would disclose my estimate of fair value for the book just to keep the analysts quiet. Like JPMorgan seems to (the details are not clear), I would also use that estimate as the basis for my loan loss provision. But then I (thank the Gods) don't work in Boadilla.

Why AIG must be kept afloat

It has written an awful lot of CDS protection. If it goes down, the contracts accelerate and the banks who bought protection get recovery on the current MTM. Recovery is unlikely to be more than 50 cents on the dollar, and so the protection buyers would suffer large losses. At least a hundred billion, I would estimate, probably more. So you have the choice between letting AIG fail and putting 100B or more into the protection buyers, or giving it another 50B and trying to keep it afloat. That's a no brainer.

The uses of common stock companies

I like Chardonnay: everything from the better Californians, like Paul Hobbs, to the majesty of mature Montrachet. But I wouldn't serve it with everything. Chocolate wouldn't work, for instance; neither would beef stew.

It is the same with stock-holder owned corporations. They are very good at a number of things, which is why they are such an important part of the financial system. But at the end of the day, a stock company exists for the good of the stock holders. It is dangerously naive to expect them to behave in the interests of any one. So it should be no surprise that the banks are finding ways to get around government restrictions on paying dividends. (The FT coverage is here.)

The big picture is that if you want to achieve an objective that is not obviously aligned with shareholder value, then using a stock company to do it might not be efficient. That holds just as much for lending as it does for running transport infrastructure or conserving the environment. It isn't that companies can't do these things. It is just that they might not be that good at it, particularly if they can't figure out how to make a lot of money in the process. If you use a flame thrower for trimming the lawn, don't be surprised when the grass is burnt.

(The building, by the way, is Lehman in London.)

Chart of the day

This, originally from Bloomberg and then picked up by FT alphaville and many others, is attracting a lot of attention:Despite some rather histrionic claims that it `proves' the bailout is inadequate and much more money is needed, I am not so sure. For one thing, banks are deleveraging, so they will need less capital. For another, various tricks are being employed, such as the FED's forebearance of capital requirements for the Bear portfolio at JPM. So the argument that the system needs as much capital as before right now is not clearly correct. Of course, capital requirements will go up eventually. But at the moment I tend towards the idea that supervisors are willing to tolerate a less well capitalised banking system at least while the crisis lasts. The banks might need more capital than is available in the bailout in due course but I am not convinced that they need it immediately.

20/20 New York Hindsight

Only four years too late, the New York Times has an article on the SEC CSE regime and how it didn't do enough to constrain the broker/dealer's risk taking. It's nice the Times is finally getting around to this kind of thing, but no one was interested in financial regulation and the scandal that was the US broker/dealer capital regime when there actually were some big broker/dealers around to take advantage of it.

Funding or capital?

I don't believe this post but I think it is interesting... John Hempton maintains that the US banking system has, or at least could soon have if it retained earnings, enough capital. [My first worry is that while this might be true on aggregate, some institutions are undercapitalised, and the market does not know who is know because they do not believe banks marks and they think there are hidden losses.]

He then maintains that the problem is funding. That is, given the banks reliance on wholesale funding, it is the inability to borrow that is sending the system to hell in a handcart - or at least sending the TED spread to 3%. Given the central bank liquidity flooding the system I don't believe this part either. I do think Hempton is right about confidence being key, but surely the cheapest way to buy some of that it to fix the underlying mortgages. Still, it is an interesting argument and I'd like to see a three way debate between John, the proponents of 'recapitalise the system' and those who think mortgage mods are the way to go.