Reg: OK Bob. You screwed up big in 08. This time, I’m gonna make sure that your bank is safe. Damn safe.

Bob: Very well, Reg, I see your point. But what exactly do you mean by safe?

Reg: I mean that you have enough capital so that you can withstand losses. Any losses your stupidity might lead you to make.

Bob: Any losses?

Reg: Yep.

Bob: What is Godzilla climbs out of the Hudson, walks down 42nd street, and eats our headquarters. At the same time all our assets, including US treasuries, fall to zero, while all of instruments we are short go up. Do we need enough capital for that?

Reg: Well, no, obviously not, that’s stupid.

Bob: Of course – no one uses 42nd street if they can avoid it. 44th would be much faster.

Reg: You’re gonna have to have enough capital to withstand a real crisis.

Bob: What kind of crisis?

Reg: A bad one.

Bob: You will forgive me if I suggest that that is less than specific.

Reg: I know your game. If I say ‘once in a hundred years’, you buy billions of once in a thousand year risk. If I say ‘once in a thousand years’, you take a leveraged position in once in ten thousand year risk. I’m not gonna get caught out like that.

Bob: So I can’t take any risk?

Reg: Well obviously you can take some. Just not too much. And we want you to keep lending to the real economy, of course.

Bob: So ordinary banking – the kind of banking that has been responsible for banks failing for hundreds of years – is fine. But anything else is to be done in moderation. How much of that is too much pray tell?

Reg: Nothing that might make me have to rescue you.

Bob: This conversation is getting a little circular…

Those who are most pessimistic about the future prospects of the borrower will be inclined to buy naked protection, while those most optimistic will be willing to sell it.

This is the first, and most important misconception. CDS trading is not about speculating on default: it is about credit spread movements. In other words, people do not buy CDS protection, typically, because they think that the market view of default is too optimistic; nor do they sell it becase they are optimists on the credit. They buy CDS because they think that the credit spread is going to go up: they sell it because they think that it is going to go down.

The next error is to suggest that protection sellers

… have to support their positions with collateral, which they do by diverting funds that would have gone to borrowers in the absence of derivatives.

This too seems to me to be a huge stretch which requires considerable justification. Collateral is usually cash, and it earns OIS. To turn liquidity into a risk position involves an asset reallocation decision which, if made generally across a bank, would dramatically change its risk profile. No, if there were no derivatives, it is highly likely that that money saved from not having to post collateral was invested in FED funds, or in some similar low risk, low return form.

Next we get to something that I think the authors should do much better. They gaily state

For reasons that are already clear from the baseline model, we find that … [some situations] involve more punitive terms for the borrower when naked credit default swaps are present than when they are not.

Given that this is the major conclusion, to say that it follows from the (in detail unspecified) model is, to put it politely, sleight of hand. If we look at the detailed paper, we find a slightly more nuanced exposition:

The availability of such contracts [as naked CDS] can shift the terms of debt contracts against borrowers by inducing optimistic investors to divert their capital away from financing real investment and towards the support of collateralized speculative positions.

In other words, the conclusions rely on the assumption that CDS protection sellers would otherwise take credit risk with their collateral, something that is highly unlikely. Typically CDS protection buyers, especially naked ones, value CDS precisely because of their liquidity and their leverage, neither feature of which is present in the bond market (unless you repo the bond – but that is another story). Thus what appears to be a carefully, academically rigourous critique of naked CDS is nothing more than a farrago of erroneous assumptions and conclusions. Caveat lector.

Update. Bug that Rajiv pointed out fixed above.

First, some general points:

• The Basel II rules are far too complex. I think I finally lost it when I got to the section on early amortisation provisions for ABCP conduits. (Hang on, I said to myself, they know about conduits – a reg cap arb device – and instead of banning them, they write rules to distinguish slightly better from slightly worse ones? You what?) Therefore I’d set an arbitrary limit, 100 pages say, and require the rules to be no longer than this, ever.
• Basel is meant to be for internationally active banks. Not for hedge funds, not for investment managers, not for corporate finance advisers. The sooner the European Commission picks up on this and implements Basel not, as currently, for tens of thousands of firms, but instead for the twenty or thirty financial institutions with more than $100B of assets in the EU, the better. Everyone else is much less likely to be systemically important, and anyway have different businesses. Write rules that suit these different types of firms: don’t impose rules designed for BofA/Deutsche/HSBC on everyone in the name of the level playing field.
• Stop fighting the last war. The number of rules or proposed rules that address what happened to AIG is absurd, for instance. You might as well ban all capital markets participants whose names begin with A and have done with it.

Given this, what can we support in Basel III?

• Capital = net tangible equity. The redefinition of regulatory capital to be based on core tier 1, or something similar, makes sense.
• Gone Concern Capital. A mechanism which would allow banks to be recapitalised by converting sub debt into equity would also clearly enhance financial stability.
• Leverage. A backstop leverage ratio provides a useful mechanism to ensure that if risk based capital rules are wrong, the resulting distortions cannot become too large. Personally I would make it inversely proportional to asset size, so the bigger you get, the lower the leverage you are required to have.

These parts of Basel III, then, are reasonable. But beyond that, what can we do with capital, and what can’t we?

• Capital requirements can make the financial system safer. There is no doubt that better capitalised firms are safer, all things being equal, than less well capitalised ones.
• However, capital is not the only tool. Indeed, there is a sense in which it is the last tool, in that if you need it, it’s a bit late. Good risk management and plentiful liquidity are vital too – and it is typically a lack of these, rather than a lack of capital, that causes firms to fail.
• Broadly, risk sensitive capital requirements are bettter. This is so obvious that it hardly needs explaining. However,once you definite a particular notion of risk, there are issues. It may be feared for instance that firms might take risk in ways that are not captured by the notion chosen, or that they might concentrate on that notion at the expense of others. The answer here is not to keep on making capital requirements more complicated until the arbitrages become hard to find, but rather for supervisors to actually understand firms’ risk taking, and to fix any obvious flaws in risk management or in capital via pillar 2.
• Thus, I wouldn’t write thousands of rules. Instead I would say something like ‘Firms must have sufficient capital to cover the losses which might be expected in a one in a hundred year financial crisis. They must be able to demonstrate that this is so, and the full details of this demonstration must be published on a quarterly basis.’ This, in Krugman’s terminology, would be a greek rather than a roman rule.
• All of this would mean that there are distortions. Bank A would have more capital for a given activity than Bank B. But that happens already – in part due to the use of internal models. But at least rather than getting false comfort from hundreds of pages of rules, supervisors, investors, counterparties and analysts would know that they had to analyse bank’s risk disclosures and capital calculations carefully.

The Options Clearing Corp. threatened to liquidate Lehman Brothers Holdings Inc.’s trading positions in the 2008 financial crisis unless Barclays Plc bought its brokerage and took on the defunct firm’s obligations, a lawyer for the U.K. bank told a judge… A similar liquidation of Lehman derivatives the same month by options and futures exchange CME Group Inc. cost Lehman’s creditors $1.2 billion, according to a report by bankruptcy examiner Anton Valukas.

Motivated by this, let me ask a hypothetical question. Suppose a firm has an asset which, were the firm to be a going concern, would produce cashflows with the PV of 100. However, the firm is not a going concern, and on forced sale, the asset is worth 60. A party approaches the liquidator and offers 80. There are no other bidders. Should the liquidator accept?

The standard theory of the duty of liquidators says yes: 80 is more than 60. This, it seems to me on a naive reading of the article (and with the usual I am not a lawyer caveats), is all that is going on here. Barclays offered less than the assets were worth as a going concern, but more than they would likely fetch in a forced sale. It’s tough for the creditors, but not as tough as a forced sale. Going concern value is not the same as forced sale value. Nothing to see here, move on.

Steadily both accounting standards and regulation has been coming around to our way of thinking (something for which we certainly claim no credit). There is a particularly clear example in the latest FSA discussion paper:

In April 2008, the Bank of England’s Financial Stability Report analysed the range of values produced by six Large Complex Financial Institutions (LCFIs) at the end of 2007 for super-senior tranches of Collateralised Debt Obligations (CDOs).

These tranches were the most senior slice of CDO structures and would therefore be expected to have a AAA credit rating at inception. The chart below shows the maximum capital requirement for such a position relative to the valuation range. In all cases, the maximum capital requirement is smaller than the variation in valuations (highest valuation minus lowest valuation reported) of the tranches produced across the six firms.

Now, it is important to understand what is claimed here. These ranges are not for the same security. One cannot infer that Bank A had a partcular mezz supersenior at 90 and Bank B had the same tranche valued at 45. All that one can infer is that Bank A had one mezz supersenior security valued 55 points over Bank B’s value for an entirely different mezz supersenior tranche. Thus much of the range reflects differences in CDO composition. Quality matters in a crisis.

The capital requirement is also misleading in that if the piece is written down from par, that loss reduces capital, so the effective capital taken is the capital requirement plus the writedown. (There is amusingly pious language elsewhere from FSA about trying to ensure that capital requirements are never more than 100% of notional – something they have thus far failed to do in various places in the capital rules.)

FSA legerdemain (which for me weakens their argument considerably) aside, there is a real point here. There are some securities whose fair value cannot always be determined accurately. It is likely that different banks will have different valuations for securities like this. Moreover, in some extreme cases, the range of values can be a significant fraction of the effective capital requirement. That’s fine, but it needs to be understood by readers of financial statements.

First she highlights an important cognitive bias:

Most of us respond better to personal stories than to impersonal numbers and ideas.

Then she discusses one of the consequences:

that same bias means we are drawn to stories about people, not systems. When it comes to the financial crisis, we want heroes and villains and what-he-had-for-breakfast narratives; we are less enthralled by analytical accounts of the global financial system and the cycle of boom and bust.

Chrystia is nice enough to suggest that this is the age of the systems thinker, that those of us who can do it – and if there is one thing that this blog is about, it is systems thinking – are the new upper class. Sadly I think she is wrong. Systems thinking has the potential to be a very powerful tool, and it has had many successes. But cognitive bias means that it is always fighting an uphill battle against personality-driven narratives. Systems thinking has a marketing problem which it needs to solve before it can become the new black.

Update. This comment from Ashwin is so pertinent that I am going to hoick it up to the main text (and edit it to remove the references).

John Sterman in his book Business Dynamics says the following: “A fundamental principle of system dynamics states that the structure of the system gives rise to its behavior. However, people have a strong tendency to attribute the behavior of others to dispositional rather than situational factors, that is, to character and especially character flaws rather than the system in which these people are acting. The tendency to blame the person rather than the system is so strong psychologists call it the “fundamental attribution error”. In complex systems, different people placed in the same structure tend to behave in similar ways. When we attribute behavior to personality we lose sight of how the structure of the system shaped our choices. The attribution of behavior to individuals and special circumstances rather than system structure diverts our attention from the high leverage points where redesigning the system or government policy can have significant, sustained, beneficial effects on performance. When we attribute behavior to people rather than system structure the focus of management becomes scapegoating and blame rather than the design of organizations in which ordinary people can achieve extraordinary results.”

Warren Buffett’s Berkshire Hathaway Inc., which has more than $60 billion at risk in derivatives, may scale back offering new contracts because of collateral- posting requirements, said a company executive…

“If you are now going to have to post dollar-for-dollar collateral, and you can’t get a price in the market that we think reflects the value of the credit quality of the company, then we wouldn’t take on that risk,” Sokol said yesterday

It is worth recapping what Buffett does in the derivatives space. He writes long dated out of the money OTC options and receives premium, which he invests. Currently he relies on the high credit quality of Berkshire Hathaway to allow him to negotiate contracts without collateral requirements. Therefore Buffett’s entire economic risk, he thinks, is the terminal value of the option: he thinks about it as an insurance contract, arguing that because he does not have to provide collateral, the path to the final value (i.e. the fair value of the option during its life) does not matter. Provided Berkshire remains well capitalised (once the fair value of the option is considered), that is true.

Now, one might argue that this was AIG’s business model in OTC derivatives too, and it didn’t work out so well for them. Fair enough. But my point is a broader one: if you require central clearing, and thus enforce collateral, this business model becomes impossible. Funding collateral is a real economic cost, and it makes Buffett’s business model much more costly and difficult to execute. In other words, one effect of requiring central clearing is that an entire business has been made essentially impossible by regulatory fiat.

I am not saying that regulators should not have the power to do that. But I do think that that power should only be used once they are sure that it is necessary for financial stability, and after considerable debate. To stop not just one firm’s activities but an entire business model just because you happen to prefer central clearing to bilateral OTC contracts seems to me to be a disproportionate reaction. How many other things are going to be ruined in this headlong rush to be seen to be doing something about financial regulation?

Investors are betting that inflation rate will rise, fattening the return on the securities and making up for the negative yield.

Or this from Seeking Alpha:

Why is the real yield negative? There are two main reasons.

1. Liquidity crisis – TIPS investors are afraid and looking for safety and liquidity at any price.
2. Inflation Expectations – The 5-Year Treasury yield is now 2.43%, so the implied inflation rate priced into the 5-Year TIPS is about 2.61% annually. This is higher than the Fed’s target inflation rate. The TIPS market is saying that the Fed is under-estimating future CPI inflation.

What’s actually going on then?

The only thing we can really deduce is that there are a lot of buyers of TIPS. What we can’t easily deduce from the TIPS yield right now is forward US inflation. Why? Well first because the argument that allows us to derive inflation from TIPS yields depends on the financing cost of nominal treasuries and linkers being the same. Right now TIPS are scarce and you cannot repo them in as easily as nominal treasuries. (This is because of retail demand, plus linkers being asset swapped to provide inflation legs against inflation swaps to pension funds.) Second, TIPS have a zero floor, so you have to option adjust the bond, but figuring out the vol to use for the option isn’t easy as there are no linkers out there without the floor and the market in inflation options is thin to say the least. (With inflation at 3% and with low inflation vol, the floor was pretty much worthless so people used to ignore it. That is no longer true.)

The first point above, then, is probably true. But don’t get carried away in thinking that the TIPS yield tells you much about inflation expectations until you have at least figured out how to account for the liquidity premium, financing and OIS effects.