New blog

I finally came up with a name that I liked for a blog and started a new one over at Wordpress titled Synthetic Assets. I just realized that I failed to post a link to the new blog from this one, so here it is.

It's not owners, but counterparties that keep banks in line

Alan Blinder has an op-ed up at the WSJ asking what to do about crazy Wall Street compensation schemes. He misses a key point: No one ever expected financial institution shareholders to be sophisticated and coordinated enough to properly monitor management decisions. The question is not why owners didn't keep managers from blowing up their firms, the question is why counterparties -- who undoubtedly understood how the game was being played and who was likely to end up eating their risky bets -- were willing to trade with ticking time bombs.

The latest straw man argument against a clearinghouse

Here's the latest argument against a clearinghouse for derivatives. From the wsj:

"It's counter to the goal of reducing systemic risk to put all the risk in one place, if that concentrates the risk of trades in clearinghouse institutions that would be 'too big to fail,'" suggests Mark Brickell, a longtime swaps banker now with an online derivatives platform called Blackbird. "It's also unnecessary. With today's ability to gather information electronically, we don't need to put all the risk in one place. We can just aggregate all the bank information in one place, so that regulators know the exposures of all the firms."

Notice what Mark Brickell is arguing: he states directly that goal is for regulators to have enough information to monitor systemic risk and presumably to use that information to impose capital controls. Earlier in the column, Gordon Crovitz argued:

Derivatives traders would have to report their trades and overall positions to regulators and could have their capital requirements adjusted. This makes sense, though regulators already had access to key derivatives information through banking regulations, with the ability to track cash flow at every bank.

In other words, Brickell and Crovitz are arguing that the goal of policy should be to have a top-down government controlled financial system even though this system has already failed us once.

Crovitz has created a straw man to knock down:

it's surprising that Treasury would create more systemic risk by putting hundreds of billions of dollars in derivatives into concentrated positions at a few clearinghouses. The rationale is that this would make it easier for regulators to gather information, but there are ways to centralize information without centralizing trading risk.

Well, no. The rationale for a clearinghouse is not to make it easier for regulators to gather information. As Crovitz has already remarked, "information" didn't help us in the past and nobody who believes in free markets would expect government regulators to be able to manage the problem of systemic risk -- especially after the current crisis.

The rationale for a clearinghouse is to create liabilities that are the joint obligation of all the members of the clearinghouse. In other words, the purpose of a clearinghouse is to place the responsibility for managing systemic risk where it belongs in a free market system -- in the hands of the banks that are in a position to create and control systemic risk.

Since a troubled clearinghouse can issue a capital call that draws resources from the member banks, the shareholders of all of the member banks can be wiped out before the government steps in to support the clearing house. In other words, while a clearinghouse is indeed “too big to fail”, because its debt is the joint responsibility of the member banks a clearinghouse is an excellent means of aligning incentives, so that regulators don’t actually need to be very attentive. The bank members of the clearinghouse will do their own policing of their fellow members, because the banks themselves will be subject to resolution if they allow the clearinghouse to get into trouble.

It hardly needs to be stated that this is precisely the kind of incentive structure to which we need to return – one where regulators can be sloppy and allow free market incentives to do their work for them.

There's another valuable piece of information that is generated by forcing market makers to jointly guarantee the markets in which they trade: Whenever the banks are unwilling to set up such a joint guarantee, it is a clear indicator that they know there is something rotten in the market. It is precisely for this reason that it is so important to force the derivatives markets into clearinghouses -- because the banks will choose to clean up the markets. The banks will dictate the terms that will make them willing to jointly guarantee the debt.

We need clearinghouses in derivatives markets, because setting them up will align incentives so that the banks choose to clean up their own market practices and do the regulators' work for them. That's what used to be called a free market economy. What Mark Brickell and Gordon Crovitz are advocating for is a socialized bastardization of the free market economy, where banks just run crying to Papa and tell him it's all his fault whenever something goes wrong.

Does Lynton Jones even know what a hybrid ABS CDO is?

There is a mind-bogglingly deceptive report issued by the City of London blaming the crisis on "over leveraged" ABS CDOs and claiming that this is evidence that credit default swaps did not lie at the heart of the crisis. This is a ridiculous point of view. Credit default swaps were the tool used by ABS CDOs to become over leveraged. While it is possible to argue that certain categories of credit default swaps are not subject to abuse (e.g. that this was a problem of index CDS and that single-name CDS are inherently safer products) or that some targeted forms of regulation are sufficient to address the problem, it is dumbfounding for someone to claim that the crisis was caused by CDOs and not by the credit default swaps they used to become overleveraged.

The confusion created by this misleading report has already led the FT to publish a commentary by someone who clearly does not understand how ABS CDOs were structured because he claims based on the City of London report: "Let’s start with the simple proposition, which seems to be almost universally accepted, that the OTC markets in credit derivatives were one of the main causes of the crisis. The City of London report utterly refutes this claim."

Why are assets toxic?

An explanation for those who can’t understand how so many assets backed by mortgages can be worth nothing at all

From 2005 through the first half of 2007 approximately $2.8 trillion in mortgage backed securities were issued by the private sector. These securities were backed by $1.1 trillion in sub-prime loans, $0.9 trillion in Alt-A loans and $0.6 trillion in prime jumbo loans (and $0.2 trillion in “other” loans). Data in intro here.

I estimate that these $2 trillion in non-prime loans and $0.6 trillion in prime jumbo loans back $4.1 trillion in AAA mortgage related securities. How is it possible to have more AAA assets than underlying loans? Through the magic of credit default swaps.

I should be clear, however: I am including in my $4.1 trillion figure $1.75 trillion of unfunded tranches of mortgage backed securities and CDOs. The data on mortgage backed securities and CDOs that is reported almost never includes unfunded tranches, because these tranches are more comparable to insurance contracts than to bonds. I include them in my estimate of triple A “assets”, because these insurance contracts can result in losses that appear on financial institution balance sheets, just as if they were funded tranches of CDOs.

Credit default swaps were used in two ways to increase the quantity of mortgage related securities available to investors. CDOs sold credit default swap protection on mortgage backed securities in order to increase their exposure to the mortgage market and CDOs bought credit default swap protection from highly rated financial institutions in lieu of selling their senior-most tranches to cash investors. (Some mortgage backed securities did the same on mortgage loans, but since the process was most commonly used by CDOs, I will refer only to CDOs from here on.)

To understand what was going on, you need to realize that mortgage securities were the hot product of the mid-naughties. Mortgages, and especially high-yield mortgages and mortgage backed securities, were in such demand that the average structured financier simply could not find enough for the CDOs he wanted to ramp up. In 2005 innovation in the credit default swap market opened up the possibility that a structured financier could create a CDO that was constructed of 10% mortgage backed securities backed by real loans (known as cash assets) and 90% credit default swaps (known as synthetic assets). Credit default swaps make a payment when the mortgage backed security referenced by the swap defaults. The investor was taking on a risk comparable to a 100% cash CDO, but the financier only needed to purchase one-tenth the quantity of mortgage backed securities he had had to find for a comparable CDO in 2004. The other nine tenths of the investors’ exposure came from using credit default swaps to reference mortgage backed securities that had already been packaged into other CDOs. Credit default swaps on mortgage backed securities made life a lot easier for structured financiers.

One consequence of the use of synthetic assets is, however, that more than one investor has exposure to each subprime mortgage backed security and thus to each underlying subprime loan. Another consequence is that structured financiers did not need cash to buy 100% of the assets in the CDO. Most credit default swaps are at least partially unfunded. This means that the seller of CDS protection promises to come up with cash in case of a default, but does not post cash to guarantee his capacity to do so when he signs the contract. When a CDO included synthetic assets, the structured financier only needed to raise enough cash – or sell enough tranches of the CDO – to cover the costs of the cash assets. You may ask, “But what’s the point of creating this huge CDO, if you’re only going to bother to fund 10% of the tranches?” The other 90% of the CDO, specifically the last-to-take-losses tranche of the CDO, could be “sold” to a large financial institution, like AIG or a monoline insurance company, or even held at the originating bank (e.g. UBS). That is, the financial institution would sell credit default protection to the CDO to cover any losses that were not covered by cash investors.

Whenever you hear someone talk about a “super senior” tranche of a security, this is precisely what they mean – some financial institution has written a credit default swap promising to pay up if defaults exceed the amount covered by the cash investors in the security. Super senior tranches of mortgage backed securities and CDOs are almost never funded.

If it happened that a structured financier had a CDO with more cash investors than cash assets, the extra cash was stored in “safe assets”, such as Treasury or agency securities or guaranteed investment contracts (GICs) issued by large insurance companies.

Now, perhaps, you begin to understand how it is possible for $2.6 trillion in mortgage loans to back $4.1 trillion in AAA securities. Quite a few of those securities were never funded and are really just promises to make a payment when there is a default on a referenced mortgage.

It’s also important to understand that the cash investors in CDOs – those who believed they were buying a “mortgage backed bond” – are going to be the first to lose their money. Why? Because there are many different classes of AAA securities – and it is the financial institutions that sold super senior credit default swap insurance that have a senior claim to any loans that do not default or that have significant recovery. The cash investors in every case bought subordinate AAA CDO securities that are guaranteed to be wiped out if defaults rise beyond a certain level. (Diagram 1 is drawn to scale, so that you can see visually the fraction of losses that will wipe out the subordinate AAA tranche.)

Notice something else about the consequences of this CDO structure given the housing crisis. As it is the super senior counterparties who have first claim to the underlying value of the mortgages, all that underlying value is likely to do is to reduce the amount that the super senior counterparties have to pay out in “insurance”. Since these counterparties initially assumed that there was almost no likelihood that they would make any payment at all, they were never prepared to pay out the whole notional value of the swap.

Thus, when spectators assume that the losses can’t really be too bad, because after all these are assets backed by real estate and there will be some recovery, they are failing to understand the nature of the ABS CDO market: (1) because the CDOs were built on high yield MBS tranches, their value can be wiped out entirely if sub-prime losses exceed 25%, Alt-A losses exceed 10% and Prime Jumbo losses exceed 3%; and (2) the super senior counterparties did not provision for making payment on their obligations. Whereas the cash investors have already put up cash that may be lost, the super senior counterparties will need to come up with the cash as defaults are recognized. Thus the super senior counterparties face not only losses on the balance sheet reducing their equity, but also a liquidity squeeze as they are forced to make payments on more than $1 trillion of swap obligations.

AIG was the first to be caught by this liquidity squeeze, but the monoline insurers are still struggling to manage the problem; the Swiss National Bank stepped in to save UBS from its embarrassments; and Citigroup had to take some super senior swaps on balance sheet in 2007 when its asset backed commercial paper conduits began to implode. Furthermore, a complete breakdown of the financial institutions exposed to super senior cash calls has yet to be made public.

Secondly, some mortgage backed securities have the same CDS based super senior structure as CDOs. While I have only located one explicit example of such an MBS (BSARM 2006-2 discussed in “Bear Stearns Quick Guide to Non-Agency Mortgage Backed Securities”) and have found no data on this issue, it was widely recognized that super senior MBS was being issued. Ominously, BSARM 2006-2 had a super senior tranche that amounted to more than 90% of the total value of the MBS. In diagram 1, I have guesstimated that $0.5 trillion of super senior swaps on MBS were issued. This would mean that in addition to the funded MBS on which we have data, there was another 20% of unfunded MBS issued.

Diagram 1 shows my estimate of the AAA assets backed by the $2.6 trillion in mortgages loans that got packaged into mortgage backed securities from 2005 through mid 2007. Diagram 2 shows the consequences of 25% losses on subprime loans, 15% losses on Alt A loans and 3% losses on prime jumbo loans.

The basic story is explained clearly in Fabozzi. My estimate of super senior ABS CDOs backed by MBS is derived as follows: From SIFMA we know that $685 billion in funded structured finance CDO tranches were issued from 2005 through 2007 Q2. I assume that 80% of the structured finance products backing these CDOs were MBS (see Fabozzi p. 175) giving us $540 billion of “mortgage backed” funded CDO tranches. I then assume that 70% of the ABS CDO structure was the unfunded super senior tranche (see Fabozzi p. 142) resulting in a total – i.e. funded plus unfunded – "mortgage backed" ABS CDO issuance of $1.8 trillion.

CDS "spread trades" and bubbles

A month or so ago there was some discussion of the "end of the world" trade. People have been trading credit default swaps (CDS) on the United States – these swaps are basically insurance against the possibility that the US defaults on its bonds – and to no one’s surprise the premium on this insurance has been going up lately. The trade in this product bewilders many for the simple reason that if the US were to default on its bonds, financial markets would be in such disarray that there is no reason to believe that any major investment bank or other seller of protection would be able to honor the insurance policy. That’s why CDS on the US are called the “end of the world” trade.

In response to the criticism of the trade, several financial commentators observed that these are in fact spread trades. (Here’s a particularly snarky example.) Because the premium is expected to go up, say from 0.7% to 0.8% over the period of a month, anyone who buys the contract today with the 0.7% premium will be due some collateral from the protection seller if the premium does in fact go up to 0.8%. (On the other hand, if the premium falls to 0.6%, the buyer of the contract will have to post collateral.)

Thus, someone who’s buying the CDS for a spread trade wants to receive collateral in case the premium on the contract goes up. One reason to do this would be to hedge: If you own Treasury bonds and are worried that interest rates will go up due to default risk and reduce the value of the bonds – the CDS can protect you from posting a mark-to-market loss on your balance sheet. But there’s a problem with this logic – interest rate swaps are a better way to protect against this possibility because they will compensate for a fall in the value of the Treasuries, whether or not the fall is due to a change in default risk. Furthermore the market in interest rate swaps is extremely deep, so the cost of entering into the interest rate swap is likely to be less than that of the CDS.

So what’s the other reason someone might buy a CDS on the US for a spread trade? To speculate on the movement of the CDS premium – if the premium moves as predicted, you can take advantage of the new premium by selling a contract at a higher premium than you are paying for a comparable insurance policy. Theoretically if there is a default requiring payment, you can simply pass on the payment that you receive from the bank that sold you protection. (Of course, this is the “end of the world” trade, so you’re probably thinking that if a default takes place, you may not get paid by the bank and you may not be able to honor your own obligations, but presumably this is such an apocalyptic scenario that it’s not really worth worrying about.)

I think the important question to ask is: What do spread trades on CDS on the US tell us about the market? They prove that there is a bubble in this CDS market.

People are trading something that they know has no fundamental value, in hopes of reaping short-term gains. Of course, the fact that the contract has no fundamental value means that someone will be caught at the top of the bubble, paying a high price for a valueless contract.

One phenomenon that supports this bubble is that, if the premium does rise, the value of the contract will go up and the buyer of the contract will benefit from short term accounting profits. The fact that everyone knows that this profit will disappear over the long-term, is a sign that these firms are “gambling for resurrection”. If they can make it through the short-term without declaring bankruptcy, they hope that longer-term investments or gambles will pay off allowing them to survive the crisis. On the other hand, they’re paying a premium today to preserve that possibility.

Who is selling the CDS on the US? Anyone with the balance sheet capacity to carry the interim losses that will have to be posted to the balance sheet if the premium goes up.

So, yes, it is true that “CDS on US government debt are spread products,” but this fact is itself a sign of how very, very sick our financial market are right now. It’s also proof positive that CDS are not always a good way to price default risk. The “end of the world” trade is a bubble and when it pops someone is sure to end up posting losses.

An invitation from the Banking Lobby: Enter the Matrix

I've been reading Congressional testimony on derivatives and I couldn't help but imagine the initial sales talk that took place some time before the action in The Matrix.

While credit derivatives are often pejoratively described in the media as a “bet”, it is important to realize that one could equally describe all investments as “bets”. When we buy the stock of a corporation, we are “betting” that the stock will be worth more in the future than what we paid.

The matrix has been carefully designed to simulate the world with which you are familiar. Any differences you perceive are not actually differences ...

The fact that the underlying credit is NOT a party to the agreement and, further, that neither the protection buyer nor the protection seller needs to own the debt of the entity, has also recently been subject to a great deal of media hyperbole. This fact is frequently and shrilly cited as evidence that credit derivatives are a “bet”. But exactly the same statement could be made about futures contracts or stock options – neither the purchaser or the seller of those contracts needs to hold a position in the underlying commodity or stock

You may be familiar with the matrix already in the form of an alternate reality games: there is really no need to distinguish between the matrix and harmless entertainment.

Consider Bank B which wants to diversify its credit exposure but does not have a relationship with the quality of credits it desires. Bank B can sell protection through a credit default swap as an alternative to making loans or buying bonds. This is economically equivalent to lending directly to the desired credits. [Read: While we don't actually lend any money, our derivatives are "economically equivalent to lending".]

When you enter the matrix we can assure you that you and your constituents will not notice that anything has changed.

It is important to remember that credit default swaps, like all derivative contracts, are zero sum contracts – the loss of one party in the contract exactly equals the gain of the other party. In aggregate, therefore, the losses incurred by protection providers equal the gains realized by protection buyers, making the overall CDS market a “closed system”, where gross losses equal gross gains, and both, when added, net to zero. This is in contrast to the cash bond market where credit losses result in permanent loss of value. [Read: Since it is unimaginable for a counterparty to go bankrupt, we have successfully created a financial product with no possibility of "permanent loss".]

Moving to the matrix brings with it great advantages. In a world without gravity, you can jump tall buildings at a single bound ...

The actual number that we should focus on is the gross replacement value of all outstanding credit default swaps, which according to the BBA was a little over $2 trillion at the end of 2007, or just under 3.5% of the notional amount for that period. That number represents the cost of replacing all the existing contracts in the market, just as the market price of an equity security represents the price at which it can be bought or sold in the open market. It is equal to the difference between the present value of fixed-rate premium payments to be made by protection buyers and the present value of the credit event-driven payments that the market expects will be made by protection sellers over the life of the swaps. [Read: Since everybody knows that we have perfected the models used to determine expected losses, you can feel confident that our prices are at least as accurate as simple stock market valuations.]

... or even dodge bullets.

In fact, according to the BBA, dealer positions represent more than 50% of the
of the credit default swap market and, as can best be determined from public
disclosures, have nearly equally balanced CDS exposures, consistent with the
dealer business model.

What? You're worried that when you and your constituents are all in the matrix, you'll be too dependent on the machines. Don't be ridiculous. As best can be determined from public disclosures, the machines are behaving consistent with our expectations. What could possibly go wrong?

The document that explains it all

From a white paper on AIGFP's employee retention plan posted on the web by Fox News:

AIGFP’s derivatives portfolio stands at about $1.6 trillion and remains a significant risk. Failure to pay the required retention payments therefore could have very significant business ramifications.

For example, AIGFP is a party to derivative and structured transactions, guaranteed by AIG, that allow counterparties to terminate in the event of a “cross default” by AIGFP or AIG. A cross default in many of these transactions is defined as a failure by AIGFP to make one or more payments in an amount that exceeds a threshold of $25 million.

In the event a counterparty elects to terminate a transaction early, such transaction will be terminated at its replacement value, less any previously posted collateral. Due to current market conditions, it is not possible to reliably estimate the replacement cost of these transactions. However, the size of the portfolio with these types of provisions is in the several hundreds of billions of dollars and a cross-default in this portfolio could trigger other cross-defaults over the entire portfolio of AIGFP.

There are also substantial risks related to the hedging of AIGFP’s various books. Although we view the large-market risk books at AIGFP as generally well hedged, the hedging is dynamic – that is, it must be monitored and adjusted continuously. To the extent that AIGFP were to lose traders who currently oversee complicated though familiar positions and know how to hedge the book, gaps in hedging could result in significant losses. This is driven to some extent by the size of the portfolios. In the interest rate book, for example, a move in market interest rates of just one basis point – that is 0.01% or one-100th of one percent – could result in a change in value of $700 million dollars if the book were not hedged. It has virtually no impact on the hedged book. There are similar exposures in the foreign exchange, commodities and equity derivatives books.

What does this mean: Bankruptcy, nationalization, the abrogation of any AIG contract worth over $25 million are all unthinkable. Why? Because the changes to the bankruptcy law that were enacted between 1982 and 2005 have privileged derivative counterparties to such an extent that any default will allow counterparties to terminate the derivative contracts.

While the laws imply that termination will involve careful netting of contracts (that is both contracts that are in the money and those that are out of the money will be terminated and then offset against each other), it is not clear to me that under ISDA protocols a derivative counterparty is required to terminate all contracts. For example, the wording above implies that AIGs counterparties have the right, but not the obligation to terminate derivative contracts. If this is indeed the case, then these counterparties are likely to argue that they have a duty to shareholders to selectively terminate those contracts that are in the money, while leaving contracts that are out of the money in place. In this circumstance the value of AIGFP's assets are likely to fall well below zero for the simple reason that AIG will be stripped of its hedges, except where they are losing money.

The preceding analysis may be wrong because I have never read the ISDA protocols. However, even if the ISDA protocols force each counterparty to either terminate all derivatives contracts with AIG or none of them in the event of a default, such a termination would strip AIG of its hedges with a potentially devastating effect on its balance sheet. (Recalling however that AIG's insurance subsidiaries may well be protected by strict regulation from the debacle.)

Why are regulators and policy-makers behaving as though firms such as AIG are holding a gun to their heads? Because the US bankruptcy law has been rewritten so that every big derivative trading financial firm is a time bomb waiting to explode. Somebody needs to defuse these bombs -- and I suspect that the only entity with that authority is Congress. The law needs to be revised so that the only circumstance in which a counterparty is granted the right to terminate derivative contracts in the event of a cross default is when this right is granted under the rules of an exchange regulated by the CFTC.

Let us not forget how these new bankruptcy laws got passed. For the most part our Congressmen and women felt that the regulation of derivative contracts was such a complex problem that only a limited group of experts could understand it. Unfortunately our regulators never really got up to speed on derivatives and also trusted the judgment of a limited group of experts. Who were these experts? Members of the financial industry itself. In particular the ISDA weighed in in support of every single one of these bankruptcy amendments.

In other words financial industry experts used their expertise to convince Congress to pass laws that gave their firms extraordinary privileges under the new bankruptcy law, in every instance resulting in a de facto taking of property from bond investors who would otherwise be senior to the derivative counterparties. Now these same experts are claiming that the government needs to go into debt to the tune of multiple trillions of dollars in order to protect the sanctity of contracts. Clearly legislative takings are okay only if they affect Main Street investors, not if they affect Wall Street.

It is time for someone to call bullshit and put an end to the extortionate scam that underlies the modern American financial industry.

Is it time to repeal the derivative related bankruptcy amendments of 1982, 1984, 1990 and 2005?

Economics of contempt states:

I'm puzzled by all the recent hyperventilating over the 2005 bankruptcy bill and the special treatment of derivatives in bankruptcy proceedings. Derivatives and certain other financial contracts are exempt from the automatic stay in bankruptcy, which essentially gives derivatives counterparties priority over all other claimants. ...

The problem with this story is that the 2005 bankruptcy bill didn't create the derivatives exemption (called a "carve-out") — the exemption for derivatives was originally enacted in 1982, and was really solidified when the statutory language was clarified in 1990. The 2005 bankruptcy bill just clarified the definitions of the various exempt contracts. For example, the pre-2005 definition of an exempt "swap agreement" included the catch-all phrase "or any other similar agreement," which almost certainly covered credit default swaps.

The last paragraph is very misleading. An exemption for derivatives was first enacted in the United States when the bankruptcy code was rewritten in 1978. The exemption was put in place so that counterparties to regulated, exchange traded futures contracts were guaranteed the right to foreclose on collateral and terminate the account of an entity that declared bankruptcy. The reason for this is that daily marking-to-market and margining (i.e. posting collateral) of positions is the mechanism exchanges use to ensure that market participants can honor their obligations. The capacity for the receiver in a bankruptcy to void margin transactions could potentially create a risk of insolvency for the exchange itself. In other words, there is indeed a genuine role for carefully constructed exceptions for certain derivatives to standard bankruptcy procedures.

What happened in 1982, 1984 and 1990 is that the financial industry argued that preferential bankruptcy treatment for exchange traded derivatives created a competitive disadvantage for over-the-counter derivatives and slowly but surely additional derivative contracts were granted safe harbor against standard provisions of the bankruptcy code too. The zeitgeist that viewed unregulated financial innovation as a purely beneficial outcome of free markets undoubtedly contributed to the steady relaxation of the law.

The argument that the 2005 bankruptcy act was just a minor adjustment to derivatives laws is simply false. The most extraordinary change took place in the area of repo contracts, but there is also a strong argument that credit default swaps were not in fact covered by earlier amendments. (A repo contract is sale and repurchase agreement, which functions just like a short-term collateralized loan.) For a thorough discussion of changes due to the 2005 bankruptcy act, see Shmuel Vasser.

The 2005 Bankruptcy Act and the repo market

According to Vasser:

Prior to the 2005 amendments, only a narrow set of securities could have been used for the repo to qualify as a safe harbor transaction. (footnote 33: These were certificates of deposit, eligible bankers' acceptances, or securities representing direct obligations of, or that are fully guaranteed by the U.S. or a U.S. agency.)

In 2005 Congress expanded the definition of a repurchase agreement eligible for safe harbor to include residential mortgage backed securities (RMBS) and derivatives on RMBS. In case you're wondering how that worked out, take a look at the table on page 11 of this paper by Gary Gorton. By July 2007 financial firms were regularly getting 100% financing against investment grade CDO collateral. One inevitable aspect of the financial crisis was the movement to more realistic haircuts on repos.

Unfortunately one of the terms used when expanding repurchase agreements to include MBS in the 2005 law was "mortgage related security." I suspect that any member of Congress who read this passage did not understand that a mortgage related security may be synthetic, that is, it may be a CDO that sells credit default swap "insurance" on a MBS and establishes some value for the CDO by positing a payout likelihood for the swap and netting this against the value of the insurance premiums. Thus a financial firm that writes an insurance policy (via a special purpose entity) was able to borrow 100% of the putative value of that "asset".

Guess what firm thought that CDO repos were a good way to finance its business: Bear Stearns. In their April 2008 congressional testimony Cox and Bernanke express absolute bewilderment about how the repo market completely dried up for Bear. Hmm, I wonder whether CDO repo markets would have grown as big and unstable as they did, if the 2005 bankruptcy act hadn't been passed and if only government and agency paper were eligible for safe harbor under the bankruptcy code.

The 2005 Bankruptcy Act and credit default swaps

Now let's talk about the extension of the derivative safe harbor provisions in the Bankruptcy code to include credit default swaps. The 1980s law that exempted interest rate swap contracts from the standard provisions of the bankruptcy code included the phrase "or any other similar agreement." The problem for financiers was that the only similarity between interest rate swaps and credit default swaps is that they both are financial contracts with the term "swap" in their names. Since common law requires courts to look at the economic substance of a transaction, there was always a strong possibility that some judge would fail to see the "similarity" between a contract that involved a repeated exchange of variable payments and a contract that was in substance an insurance contract, where the payments made by the two parties were separated over time. Thus, it was only with the passage of the 2005 bankruptcy act that a financial market participant could be confident that credit default swaps were exempt from the standard provisions of the bankruptcy code.

The 2005 Bankruptcy Act and systemic risk

Needless to say, plenty of people realized that giving special privileges under bankruptcy law to over the counter derivatives is just as likely to increase systemic risk as to reduce it.

Economists Robert Bliss and George Kaufman:

We conclude that the systemic risk reduction claims often made for close-out
netting and collateral protection appear at a minimum to have been over stated. Systemic risk is in part made more likely as a result of these protections, but then so also are the benefits obtained from a more efficient market that is based on these same protections. The combined use of these three provisions represent a two-edged sword that cuts both ways.

Bliss and Kaufman make the point that should have been obvious to everybody: When you reduce the likelihood that an individual firm will face the costs of counterparty risk, you may find that you increase the size of the market so much that there is no reduction in counterparty risk in the aggregate.

Law professors Franklin Edwards and Edward Morrison had a different concern:

as the LTCM experience demonstrates, permitting the immediate liquidation of a large financial institution counterparty such as LTCM can generate another form of systemic risk, namely the risk that a a run by derivatives counterparties on the debtor will itself destabilize financial markets.

This appears to describe precisely what happened when Lehman Brothers failed.

Finally, we have Shmuel Vasser again:

[I]n many derivative transactions the bankruptcy filing merely subjects the non-debtor counterparty ... to a risk of loss, not an uncommon situation for parties with business relationships with a failed enterprise. The safe harbor provisions change all of that.

Assume that a debtor uses large amounts of metal in its manufacturing process and to protect itself from price fluctuations has entered into a long term forward for its metal requirements. At the time of its bankruptcy, metal prices are significantly higher thus making the forward a highly valuable estate asset, and potentially making its reorganization more likely. Not so; upon bankruptcy the seller can terminate the forward allowing it to obtain market prices for the metal. If, however, at the time of bankruptcy metal prices are lower than the contract price, the non-debtor counterparty has the option not to terminate the forward, and file rejection damage claim should the debtor decide to reject the forward. As this example makes clear, only the non-debtor counterparty obtains the upside of a derivative in a bankruptcy, not the debtor.

... Now assume, however, that the debtor in the example above is a major corporation with tens of thousands of employees, tens of billions of dollars in debt securities outstanding, millions of public shareholders including pension funds and billions of dollars of notional amount of derivatives -- will the application of the safe harbor provisions in such a case serve to protect financial markets from systemic risk or stress them even further? Time will tell.

Vasser points out that the "safe harbor" provisions for derivatives in the bankruptcy code will result in a bankrupt firm being stripped by counterparties of its hedges and thus preclude a Chapter 11 reorganization of the firm. The fact that a financial firm will be forced to declare Chapter 7 bankruptcy (that is to liquidate) is common knowledge. What is not common knowledge is the fact that stripping a financial firm of all its hedges before liquidation is very likely to leave the firm with almost no value for bondholders to recover.

Thus, the derivative related revisions to the bankruptcy code of 1982, 1984, 1990 and their final apotheosis in 2005 created a huge class of secured creditors which has priority in bankruptcy over unsecured creditors (i.e. bondholders). Effectively bonds in financial institutions (and other firms with significant derivative exposure) were converted by revisions to the bankruptcy law into a senior form of equity. When this fact is fully understood, it is very likely financial firms will no longer be able to raise money on commercial paper and bond markets.

In short, those who argue that the government needs to take action X, Y or Z in order for financial firms to fund themselves on private capital or debt markets are ignoring the underlying problem. Until the special privileges granted to over the counter derivatives contracts are repealed by Congress, there is no reason to believe that financial firms will ever be able to raise money on capital or debt markets again in the absence of a government guarantee – for the simple reason that once derivative counterparties are done with a bankrupt financial firm there is not likely to be much left for residual claimants.

This is the true lesson to be drawn from the failure of Lehman Brothers and the constant refrain “No More Lehmans”: It may be time to repeal some of the special privileges granted to over the counter derivatives in the 1982, 1984, 1990 and 2005 bankruptcy law amendments.

How bad is public-private investment?

The Obama administration is taking some flak from economists for its proposal to finance private purchases of toxic assets using non-recourse loans. I, however, can understand the logic behind the proposal. If you believe that the collapse in credit is overshooting, so that the shrinking of the intermediated money supply is more than would be required in a normal environment, then this policy makes sense. (It seems to me that everyone who supports fiscal stimulus is implicitly taking this point of view.) In short there is some reasonable leverage ratio for non-bank financial institutions where the asset to equity ratio lies, I would estimate, between 4 and 10. If non-bank financial firms are being forced to a leverage ratio lower than this, then an unreasonable amount of leverage is being squeezed out of the financial system, and we are failing to use our savings efficiently. In other words, it is possible that some well-managed hedge funds (et al.) are in fact underleveraged right now. For this reason, the government may choose to step in to support asset values by deliberately increasing the amount of leverage in the financial system.

While this is certainly a somewhat risky position for the government to take, if the policy is implemented with extreme care, it may be successful. Here are the caveats I would propose to help establish "market" prices in our current not-so-market financial world.

(i) The fraction of money that private funds put up needs to be high (25% sounds about right to me) and that private money should be in a first-to-lose, last-to-gain position -- in other words, all income and principle needs to go to the government before the private investors get a penny. Furthermore, there need to be safeguards in place to ensure that the private stake is real. (Maybe, the crisis has made me cynical, but I genuinely wonder whether its possible for the government to be tricked into putting up close to 100% of the funding.)

(ii) No synthetic assets should be funded using this mechanism. If the government wants to support the purchase of synthetics, it should do so only on the basis of full recourse loans to well-capitalized entities. The reason for this is simple: As a matter of market economics, one should assume that on Sept. 17, the day AIG was bailed out, there was a total domino-like collapse of the investment banks and any bank obligation that is not backed by real collateral was left with nothing more than the option value of government intervention. Thus, on Sept. 17 the intrinsic value of all CDS contracts should be assumed to have fallen to the amount of collateral that was posted as of that date.

If the government starts offering non-recourse loans for the purchase of assets that have value only because of the option value of government action, it will be an exercise in the absurd. Effectively the only reason to buy synthetic assets is because the purchaser believes that the government will be so embarrassed by the losses on the nonrecourse loans that will be incurred by a government failure to support the value of CDS that the government will engage in a costly taxpayer funded bailout in order to cover up its own mistakes. Trust me, we don't want to go there.

Update 3-4-09: I reread the post several times yesterday asking myself whether it wasn't maybe a little over the top, but decided to leave it alone. And then today I find that El-Erian agrees with me: “You can no longer predict asset value without thinking about the role of governments. Governments are no longer referees, they are players.”

Principles for future financial regulation (updated)

This post by Stephany Griffith-Jones states: “There are two broad principles on which future financial regulation needs to be built.” I agreed with this statement and read with curiosity her principles. I don't think her first basic principle is sufficient. (Her second is pretty similar to my second principle below.) She proposes countercyclical capital requirements. Like the "liquidity charges" proposed here, I think these authors are over-optimistic about the capacity of regulators to make good decisions about things like the price of insurance and the correct dynamic level of capital.

The environment in which regulators operate needs to be remembered: the entities with the most time and energy to communicate with regulators are the regulated firms themselves. This creates an inherent tendency in regulators to under-regulate especially when the economy is stable over a long period of time. The fact that in a crisis the regulators will be blamed for capital requirements that are too low or for under pricing insurance works to absolve financial institutions for their own failures in a way that is totally inappropriate.

In short, the incentives in these schemes are all wrong. When crises occur and financial institutions fail, we need to have a regulatory structure that makes it clear that these failures are completely and entirely the fault of the financial institutions themselves -- for the simple reason that the ones with the greatest ability to prevent the crises are the financial institutions.

I would propose one meta-principle from which two subsequent principles follow:

I. Competition without bankruptcy and failure is not competition.

This principle implies:

(i) All financial institutions that carry government guarantees should be protected from so-called market competition. Banks with deposit insurance need to be considered public utilities: They can earn a rate of return deemed fair by regulators, but will be safe investments with a stable, but never remarkable return. If money market funds are to be brought permanently under the federal safety net, then they will be treated similarly. In this case, we need to recognize that the terms of the “competition” between money market funds and banks will be set by government policy, not by the market.

The logic behind this view is simple: there are certain core functions of the financial system that the government chooses to guarantee. One of these is bank deposits. As long as bank deposits constitute a core function guaranteed by the government, any close substitute for that core function will also end up (possibly ex post) with a guarantee. Example: money market funds.

Thus, it makes no sense for there to be competition in guaranteed core functions of the financial system. The financial sector may be permitted to develop such alternatives, but as soon as they are large enough to compete with the guaranteed core function, the playing field must be evened by charging them guarantee fees and bringing them into the protected financial sector.

(ii) Any financial institutions that are left open to competitive market forces must be allowed to fail on a regular basis. In order to make the commitment to failure credible, the government will probably have to regulate the size of these financial institutions individually, the extent of their liabilities to the protected financial sector, the extent of their (conditional) claims on the protected financial sector and the size of the markets they create in the aggregate. Markets that grow large will have to move to regulated exchanges, so they can be carefully monitored to ensure that they carry no risks for the protected financial sector.

A riff on Knightian uncertainty and asset pricing

jck at aleablog quotes Donald MacKenzie on the Single-Factor Gaussian Copula Model or why simple models survive:

The availability of conceptual equipment can matter even if the theory underpinning the equipment is not understood -software systems allow traders with only a rough grasp of the theory of options or of CDOs to calculate implied volatilities or base correlations- or not believed. Those who do understand the models that are used in such calculations frequently view them as oversimplifications. I have, for example, yet to interview a credit-derivatives trader who regards as adequate the ’single-factor Gaussian copula’ model normally used in credit correlation calculations. Nevertheless, the simple models remain in wide use. More complex models face formidable barriers as communicative tools, because for full communication both parties must be using the same model, and that is seldom the case once one moves beyond simple models. Furthermore, the simple models typically have just one free parameter -’implied volatility’, for example- with the other parameters being either fixed by market convention (CDO pricing, for example, was often done assuming a recovery rate after default of 40 per cent, whatever the corpoaration that has issued the debt in question) or regarded as empirically observed facts. When numbers of free parameters are larger, or parameters do not have intuitive interpretations -as is often the case with more complex models- communication and negotiation become much harder.

This is how I understand this quote: The act of pricing complex instruments in financial markets requires addressing Knightian uncertainty -- or deliberately choosing to ignore its effects on asset performance by using a model that everyone recognizes is oversimplified.

But why would anyone choose to trade an asset that they know is being priced incorrectly? Maybe this is what the broker-dealers are thinking: As long as we can get profits by brokering this stuff, why should we bother worrying about the accuracy of the prices we are putting on our products?

In addition to fees, another advantage of a broker dealer keeping a mispriced market going is that in the instances where the mispricing is so bad that the dealer knows for sure which way the market is mispriced, there's a profit opportunity too. If we assume that arbitrage takes place fast enough then we can assume that when dealers are pricing with oversimplified models, the prices will be kept within the bounds defined by Knightian uncertainty -- if prices get too far out of whack dealers will step in to correct the mispricing. As long as the mispricing lies within the bounds of Knightian uncertainty, dealers will be content to broker the mispriced product to their clients.

Note: jck also links to an excellent Mackenzie article from last year.

Understanding what bankers mean by Knightian uncertainty

In previous posts I have discussed the claim that there has been a sudden increase in Knightian uncertainty since the onset of the crisis. Andrew Haldane of the Bank of England has written a paper that helps explain the origins of this claim:

Each of these market failures has been exposed by events over the past 18 months. When risks materialised outside of calibrated distributions, risk models provided little guidance in identifying, pricing and hence managing them. This failure is not of purely academic interest. The breakdown of risk models is itself likely to have contributed importantly to crisis dynamics. Why?

First, the potential losses arising from under-pricing of risk are large. …

Second, the breakdown of these models had the consequence of turning risk into uncertainty, in the Knightian sense.11 Once the models broke down, how were assets to be priced? Practitioners have a devil of a job pricing assets in the face of such uncertainty. So too do academics, though some attempts have been made.12 The theory of asset pricing under Knightian uncertainty throws up at least two striking results. First, in the face of such uncertainty, asset prices are not precisely determined but instead lie in a range. This indeterminacy in prices is larger the greater is uncertainty and the greater agents’ aversion to it. Second, asset prices exhibit a downward bias relative to fundamentals. Uncertainty gives the appearance of “pessimistic” expectations.

Apparently the reason for this abrupt outbreak of Knightian uncertainty is that bankers have suddenly realized that there is a difference between reality and their models. As long as the world behaves according to model, bankers want to claim that they are earning profits from managing “risk,” and as soon as their models fail, risk becomes uncertainty and necessitates a government bailout.

In short whenever you read that bankers can manage risk, but uncertainty requires government intervention, you should hear: “Privatize the profits and socialize the losses.”

The truth is that bankers always have to price assets in the face of Knightian uncertainty, they have just chosen to spend the last decade pretending that this was not their job.


Note: The difference between Haldane's approach and mine is that Haldane believes that is possible to change the models so that they will be "roughly right," whereas I believe that the fundamental flaw was the belief that any model could be "roughly right." Don't get me wrong. I'm not a Luddite. I don't think we should stop analyzing data using models. I just think that every time you use a model, you need to be aware of the many ways in which it is totally and completely wrong. In short, models are never intrinsically valuable, but careful and constructive human interaction with models can be priceless. Update 2-25-09: Have finally read Haldane with some care. This may not be so much of a difference. Haldane's just talking like a regulator who has to design concrete tests for banks to perform. He might agree that the ideal is to keep updating those tests regularly as regulators' understanding and experience changes (and to keep the banks on their toes).


Update 2-15-09: There is a problem with Haldane's use of the term Knightian uncertainty (and probably with the bankers' use of it). Haldane's paper argues that bankers measured uncertainty incorrectly. He then states "the breakdown of these models had the consequence of turning risk into uncertainty, in the Knightian sense."

This statement confuses Knightian or unmeasurable uncertainty with unmeasured uncertainty. The bankers' models were wrong, they mismeasured risk. The bankers apparently wish to claim that this unmeasured uncertainty was unmeasurable -- that is why they call it Knightian. Haldane by contrast argues that this unmeasured uncertainty is in fact measurable by using a broader range of data and stress testing.

In short, when bankers claim that Knightian uncertainty has increased, they are just trying to make excuses for their own failures. Haldane has recognized the banker's failure to properly model risk, and doesn't appear to be using the term "Knightian uncertainty" in a manner consistent with its original meaning.

Has uncertainty increased?

I was going to put a snarky comment up at Free Exchange in response to the post on economists being absolved by the sheer irrationality of recent market events, but didn't because I thought just maybe there was in fact evidence of an increase in Knightian uncertainty or "unknown unknowns."

The basic argument is this: The market is panicked and irrational. In the face of Knightian uncertainty, consumers and firms are refusing to take on profitable risky opportunities. This has led to a massive market failure.

I guess I'm just too much of a believer in markets to fall for this. It looks to me like we have an overleveraged economy that has experienced the shock that will force it to deleverage. This means that mean economic outlook for both consumers and producers is much worse than it was six months ago. In view of the poor economic outlook, firms are less willing to invest, while consumers are saving more and storing their savings in places where they're pretty sure the principal won't disappear. There's no need for uncertainty to generate either a bad outcome or, in the absence of aggressive policy, a downward spiral.

And then I realized that maybe there really is very little evidence of an increase in Knightian uncertainty. Remember Knightian uncertainty is the one we can't measure (i.e. it's not something as simple as the risk of default of an single firm), so it's pretty much an oxymoron to claim that we have demonstrable evidence of an increase in Knightian uncertainty.

Not only that but today I read Lloyd Blankfein in the FT: Since the spring, and most acutely this autumn, a global contagion of fear and panic has choked off the arteries of finance, compounding a broader deterioration in the global economy. Looks to me like the bankers are going to milk that uncertainty model for every penny they can get out of it.

So I think it's time to talk about some common measures of economic uncertainty that are used: the CDX-IG, or the market's measure of the likelihood that investment grade firms will default, and the VIX, or expected stock market volatility implied by the pricing of options. The basic problem with claiming an increase in these indices as evidence of an increase in Knightian uncertainty is that both of these are simple measures of risk -- they measure known unknowns, not unknown unknowns.

Thus the rise in the CDX-IG is just an indicator that expected corporate default rates have increased. This is a pretty new index, but it is surely perfectly predictable that in every recession the CDX-IG would go up. If it's a good measure of risk, it should probably start to go up before the recession starts and to go down before the recession ends, but cyclicality would simply reflect the normal operation of the economy.

Furthermore a simple review of the VIX formula will show that it's guaranteed to rise markedly when the stock market falls by half. Taking this increase as a sign of Knightian uncertainty is foolish.

Perhaps one would argue that the volatility of the CDX-IG has risen and this is a sign of an increase in Knightian uncertainty. However, the CDX-IG is a very thinly traded contract: According to the DTCC through all of the week ending January 23, there were less than 1700 position increases and 700 position decreases total in the four different vintages of the Markit CDX-IG that traded that week. (Trade in the Dow Jones CDX-IG is much smaller.) In the meanwhile on January 23 there were 16 market makers with on average just over 2,800 contracts selling protection each and on average just under 3,000 contracts buying protection each. In a market with this kind of structure, a large dealer that needs to -- or chooses to -- adjust its position is likely to have a "technical" effect on the market.

That prices would be volatile in thinly traded markets is not surprising. The real question that an increase in volatility raises is: What was going on in the two years preceding the outbreak of the crisis to repress the volatility of the CDX IG? My guess is that that financial innovation was steadily bringing new sellers of credit default protection into the market (such as school districts and Australian towns). As soon as the field of sellers stopped expanding -- and in fact shrank as schools and towns realized how much risk they were taking on -- the CDX-IG rose and its volatility returned to a normal level.

19th v 21st c. banking: What has changed ...

There's one last quote from Lombard Street that I want to record. In reference to the Bank of England's lender of last resort activities, Bagehot writes:

No advances indeed need to made by with the Bank will ultimately lose. The amount of bad business in commercial countries is an infinitesimal small fraction of the whole business. That in a panic the bank, or banks, holding the ultimate reserve should refuse bad bills or bad securities will not make the panic really worse; the 'unsound' people are a feeble minority, and they are afraid even to look frightened for fear their unsoundness may be detected.

Is anybody else wondering whether it's still true that: "the amount of bad business in commercial countries is an infinitesimal small fraction of the whole"?

Why Bagehot wrote Lombard Street

A recent exchange made me think for the first time about the events preceding the publication of Lombard Street. The financial crisis of 1866 had rocked the financial system dramatically without really destabilizing it. But Bagehot saw in that crisis the possibility of total financial collapse. (A detailed description of financial collapse takes up much of Chapter 7, part II of Lombard Street.) For this reason he thought it was important to explain what made it possible for the British economy to survive such brutal storms -- and thus to make it clear to politicians and to the Governors of the Bank of England what their obligations were in a crisis.

After 1844 politicians played an important role in the Bank of England's lender of last resort activities. The Banking Act of 1844 (Peel's Act) restricted the issue of Bank Notes. It was, however, written with an escape clause, allowing the law to be suspended by executive order (i.e. bypassing Parliament). And in every subsequent crisis 1847, 1857, 1866, the law had to be suspended in order for the Bank of England to act as lender of last resort to the banking system. That is, without suspension of the law the Bank would have run out of Bank Notes.

Peel's Act was so controversial that Bagehot states: "Two hosts of eager disputants on this subject ask of every new writer the one question -- Are you with us or against us? and they care for little else." For this reason Bagehot chooses not to take sides in the debate, but instead to explain clearly (without condemning the law itself) why Peel's Act had to be suspended in 1847, 1857 and 1866.

Bagehot's main concern in Lombard Street is to put an end to any controversy over:
(i) the importance of suspending Peel's Act in a financial crisis
(ii) the importance of copious lending by the Bank of England in a crisis.
Chapter 7 is dedicated to explaining that public uncertainty about either of these two actions in a crisis is dangerous. He concludes:

The best palliative to a panic is a confidence in the adequate amount of the Bank reserve, and in the efficient use of that reserve. And until we have on this point a clear understanding with the Bank of England, both our liability to crises and our terror at crises will always be greater than they would otherwise be.

Note the subtle reference to Peel's Act in the first sentence. Bagehot avoids controversy, but makes his point clear. In the second sentence he calls on the Bank to publicly acknowledge its role as lender of last resort to the financial system.

One thing needs to be emphasized: Bagehot is concerned about uncertainty -- but that uncertainty references explicitly whether or not a lender of last resort exists and will step in to provide liquidity in a crisis. Thus Bagehot wants the Bank of England to reduce uncertainty in the financial system, but only by promising to lend copiously against high quality assets.

He is entirely at ease with the fact that the Bank of England allowed Overends Gurney, a systemically important bill broker, to fail despite the fact that the failure rocked the financial system with uncertainty, as illustrated by the following two quotes:

In 1866 undoubtedly a panic occurred, but I do not think that the Bank of England can be blamed for it. They had in their till an exceedingly good reserve according to the estimate of that time—a sufficient reserve, in all probability, to have coped with the crises of 1847 and 1857. The suspension of Overend and Gurney—the most trusted private firm in England—caused an alarm, in suddenness and magnitude, without example.

no cause is more capable of producing a panic, perhaps none is so capable, as the failure of a first-rate joint stock bank in London. Such an event would have something like the effect of the failure of Overend, Gurney and Co.; scarcely any other event would have an equal effect.

Thus, Bagehot does not argue that the lender of last resort should act to eliminate general uncertainty, he argues only that a lender of last resort should eliminate uncertainty about the actions it will take in the midst of a financial panic.

Learning from the Crises of 1857 and 1866 Continued

In my last post I never did get to the parallels between 1857/66 and today. First of all note that in 1857 Gurneys was a huge player in the money market and thus in 1857 when the Bank of England lent £9 million to the bill brokers, you can expect that at least £2 million and possibly as much as £4 million went to Gurneys.

In other words the Bank of England probably saw evidence that in a crisis the line of credit it was giving to an individual firm was becoming unreasonably large. Even if all the bills discounted at the Bank by Gurneys were good quality bills at this particular time, the Bank was opening itself to a future problem where reliance on the careful management of Gurneys allowed low quality bills to be discounted at the Bank and thus exposed the Bank to significant credit losses. As the Bank was a private institution, it had an obvious interest in limiting it's exposure to the credit risk of a single firm or to the credit judgment of a single underwriter of commercial bills.

In fact, it was precisely the ultra-conservative management of the Bank of England that created the trust necessary for the Bank to be relied on as a lender of last resort:

The great respectability of the directors, and the steady attention many of them have always given the business of the Bank, have kept it entirely free from anything dishonorable and discreditable. Steady merchants collected in council are an admirable judge of bills and securities. They always know the questionable standing of dangerous persons; they are quick to note the smallest signs of corrupt transactions; and no sophistry will persuade the best of them out of their good instincts. You could not have made the directors of the Bank of England do the sort of business which 'Overends' at last did, except by a moral miracle—except by changing their nature. And the fatal career of the Bank of the United States would, under their management, have been equally impossible. Of the ultimate solvency of the Bank of England, or of the eventual safety of its vast capital, even at the worst periods of its history, there has not been the least doubt.

Thus, in 1858 the Bank put in place a policy that only allowed bill brokers access to emergency credit by exception. The purpose of the policy was to force bill brokers to keep their own reserve, or in other words to be prepared to act as their own lenders of last resort.

Thus, faced with financial innovation and the growth of an extremely large, highly leveraged firm closely intertwined with the banking system, the Bank of England immediately made it clear that it did not support the growth of this firm. It set a policy: if financial innovation was going to allow huge firms to develop, then those firms should be prepared to support themselves through a crisis.

One obvious consequence of this policy is to push the bill brokers to reduce their leverage ratios and thus reduce the likelihood that they will (i) need the services of a lender of last resort and (ii) cause a financial crisis by failing. In short, even though the Bank of England had no regulatory authority whatsoever over the financial system, the fact that it was the lender of last resort meant that its policy decisions affected the stability of the financial system by affecting the leverage in the financial system.

Contrast the behavior of the Bank of England in 1857-8 with the behavior of authorities in the US after the 1987 investment banking crisis. In 1987 the stock market crash could have led to the complete collapse of the investment banking industry were it not for the actions of the New York Federal Reserve President (who basically told the commercial banks to give the investment banks unsecured loans). Notice that at the moment of the 1857 and the 1987 crises both central banks took aggressive action. The difference is in their behavior after the crisis was over. The Bank of England immediately acted in a manner that would push the firms that were endangering the financial system to reduce their leverage ratios and be less reliant on the Bank of England in the future. By contrast, the Federal Reserve stood by without objection when Congress extended it's authority to permit it to lend directly to the investment banks in the 1991 FDICIA law.

In other words the two Banks took diametrically opposed actions: that of the Bank of England would tend over the long run to decrease the leverage of financial institutions and thus decrease the risk of financial instability, whereas that of the Fed tended to increase leverage and increase instability over the long run. Observe, however, that the short run effects of these different policies were precisely the opposite of their long run effects.

The willingness of the Bank of England to let a bill broker fail was tested within a decade. Overend and Gurneys actually made the decision easier by being the subject of rumors (which turned out to be well founded) for over a year before they finally turned to the Bank of England for aid in 1866. (The Economist wrote at the time "the failure of Overend, Gurney and Co. Ltd. has given rise to a panic more suitable to their historical than to their recent reputation." Victor Morgan, 1943) Despite their size, the Bank of England did indeed refuse to discount their bills -- and the result was a financial crisis that matched the worst in living memory.

On the other hand, no further failures threatened to rock the market until 1890, when Barings' exposure to South American loans could have resulted in failure and a major crisis. The potential losses were so great that the Bank of England refused to intervene directly, but it did broker a joint support from the other large banking houses. Overall, however, England's financial system was remarkably stable from 1866 up through 1914 (the year which marked the beginning of the end of the gold standard).

By contrast the Federal Reserve's policy of bringing the investment banks under their wing forced the Fed in 1998 to extend the safety net even further. When the Fed brokered a bailout, it was not a bailout of a systemically important bank, it was the bailout of a hedge fund. In hindsight it's extremely easy to see that the fact that a hedge fund had grown large enough to be systemically important was a clear sign that the financial system was dangerously overleveraged.

If the Fed had been a private bank like the Bank of England in 1857, at this juncture it would have made it clear to all members of the financial system that its balance sheet could not support the liabilities that were growing. It would have told them that they needed to start paying attention to their own reserves by dramatically reducing their leverage ratios.

Instead, the calming effects of an endless sequence of government bailouts of the banking system (Continental Illinois in 1984, Brady Bonds in 1989, low interest rates to ease the burden of bad debts in the early 90s, the late 90s and early noughts) left the Fed with the illusion that financial instability was no longer a problem that needed to be addressed, while at the same time fostering an accumulation of dead wood in the financial system that ensured that the resulting conflagration would be beyond all imagination.

In short, while the Bank of England in 1858 was willing to precipitate a crisis in order to foster stability, the Fed in the 1980s and 90s fostered the illusion of stability, while at the same time creating an environment where crisis was inevitable. Thus, just as the forest manager is better off allowing small fires to flare up regularly, so a central banker is better off precipitating small crises to avoid a disastrous and uncontrolled burn.

Learning from the Crises of 1857 and 1866

I've been rereading Lombard Street, this time reading it as a history of the money market in mid-nineteenth century England. There are some very, very interesting parallels to the current crisis.

In particular, the period (after the passing of the Bank Charter of 1844) was one of remarkable financial innovation. On the one hand, banks that had initially issued circulating notes adapted to the new environment by converting to deposit banks and eventually offering checking accounts. On the other hand, new bank like financial intermediaries were developing. Bagehot calls these financial intermediaries bill brokers, and elsewhere they are called discount houses. In any case, as Bagehot makes clear bill brokers, that in the past were pure brokers matching savers with credit-worthy borrowers, had become intermediaries who guaranteed the bills that they placed in large quantities with banks. This was a tight margin business where the "brokers" borrowed most of their capital from these same banks and thus they were essentially earning money on spreads and the value of their highly specialized knowledge of the quality of commercial bills. Aggressive competition meant that it was not profitable for these brokers to maintain reserves (i.e. capital) to back up their guarantees. Instead the bill brokers relied on the Bank of England's discount policy: Even in the worst of crises the Bank was expected to discount good bills at Bank rate.

In the crisis of 1857, the Bank of England advanced more than £9 million to the bill brokers and only £8 million to bankers. After the crisis, the Bank stated a new policy which restricted the Bank tranactions of the bill brokers to advances which were only offered on a quarterly basis. Should the brokers need discounts at any other time, they would have to ask for an exception to the Bank's policy. The stated goal of the policy was to make the brokers "keep their own reserve."

Needless to say, the bill brokers were unhappy with the new policy. Gurneys, which controlled more than half of the commercial bill market and was in many ways a competitor of the Bank of England, apparently tried to start a run on the Bank in April of 1860 by withdrawing from its deposit account an outrageous sum of money all at once. (While Bagehot claims the sum was £3 million, other sources cite a figure of £1.65 million.) This action was roundly condemned in the press.

In the meanwhile Gurneys was being poorly run by a second generation of the family. As badly underwritten bills went into default, the company advanced money on mortgages and even ended up running a fleet of steamships -- also unsuccessfully. By matching short term obligations with long term assets, Gurneys violated one of the most fundamental principles of 19th century banking. This explains the strong words Bagehot uses to describe the company:

The case of Overend, Gurney and Co., the model instance of all evil in business, is a most alarming example of [the] evil [of a hereditary business of great magnitude]. No cleverer men of business probably ... could well be found than the founders and first managers of that house. But in a very few years the rule in it passed to a generation whose folly surpassed the usual limit of imaginable incapacity. In a short time they substituted ruin for prosperity and changed opulence into insolvency.

In 1865 before all of Gurney's problems were public knowledge, the partners took the firm public, creating Overend, Gurney and Co. While this action required the Gurney family to guarantee the new firm against losses on the business of the old, presumably it was hoped that new capital would save the firm. Unfortunately the losses were so great that the Gurneys had to liquidate their personal property and news of this event caused a run on the Company. Those who had bought shares in 1865 ended up losing £2.9 million. (Because they had only paid 30% of the face value of their shares, they had the misfortune to face a capital call in order to satisfy obligations to creditors -- who were paid in full.) The lawsuit that followed this failure found the Gurneys' actions to be incompetent rather than fraudulent and they were acquitted. (Ackrill and Hannah, 2001, Barclays: The Business of Banking, 1690-1996, p. 46-7)

Because many London banks were exposed to Gurneys, there was a run on the London banks and many solvent banks, such as the Bank of London, failed. Wikipedia claims that in the crisis that followed there were a total of 200 bank and commercial failures. By any standard the crisis was severe. In particular because London held large foreign deposits that were slow to return after the run, the Bank of England was forced to keep the Bank Rate elevated for a full three months -- which undoubtedly aggravated the domestic consequences of the crisis.

It was, however, a watershed because for the first time the Governor of the Bank of England publicly acknowledged "a duty ... of supporting the banking community", that is, he acknowledged that the bank was the lender of last resort to the banking system. (While the Bank had been playing this role for almost a century, it often did so with reluctance and heretofore had never publicly recognized the role as an obligation.)

For this reason, Bagehot repeatedly treats the Bank of England's actions in 1866 as the model for a lender of last resort. A few points are worth mentioning:

(i) In Bagehot's time, it was exceptional for lender of last resort activities to last more than a few weeks. Frequently the simple fact that they were available (e.g. a lifting of government restrictions) was enough to end the panic.

(ii) A lender of last resort is not expected to prevent the failure of a systemically important bank. On the contrary, Overend, Gurney and Co. was systemically important, but it was also so badly managed that the Bank of England could not be expected to discount its paper.

(iii) A successful lender of last resort action will leave the bulk of the financial system standing (i.e. at least say 75% to 90% of the banks). Bank failures -- even large numbers of bank failures -- are part of a typical lender of last resort activity.

(iv) The government should never support a bad bank; such action could only serve to prevent the development of good banks.

So long as the security of the Money Market is not entirely to be relied on, the Goverment of a country had much better leave it to itself and keep its own money. If the banks are bad, they will certainly continue bad and will probably become worse if the Government sustains and encourages them. The cardinal maxim is, that any aid to a present bad Bank is the surest mode of preventing the establishment of a future good Bank.

Update: This post is continued here.

Solutions to the CDS overhang

Solution 1: My first thought was this: Since too many of the dealer-counterparties can't honor their commitments and a chain of failures is the necessary consequence, there's nothing to do but nationalize all the dealer banks and, by aggregating them, wipe out their commitments to one another. However, since JP Morgan seems to have been relatively conservatively run, this is somewhat unfortunate. I imagined a conversation between Geithner and Dimon as follows: "Look, you prepared for a crisis, just not this crisis. I'm sorry we've got to nationalize JPM, but we do have a consolation prize: How would like to manage the great stinking bankpile as your government portfolio?"
(Look, Ma, bank's a four-letter word. Hoocoodanode?)

Solution 2: But then I went for my weekend jog and came up with what I think is a much better solution. Before buying or guaranteeing toxic assets, the government should require participation fees from the banks on the following basis: For each participating bank the Fed should sum the absolute value of the bank's net notional positions over each category of CDS. Let's call this number the bank's non-dealer position. The participation fee should be the non-dealer position (possibly after subtracting off the lowest non-dealer position, so that one dealer bank is guaranteed to pay nothing). The government may choose to finance this fee as a preferred stock-type loan.

What is the advantage of the non-dealer position fee? Two types of banks are creating chaos in the markets: those that guaranteed debt they could not afford to guarantee and those that ignored the dangerous exposure of their counterparties and bought insurance that they knew (or should have known) could only be honored if the government intervened in the market. Both sides of these trades were toxic. On the other hand, any bank that actually behaved as a dealer and thus carefully maintained offsetting positions in each market will find that its non-dealer position is small. This latter type of bank was relying only on the market continuing to function, not on being able to extract money from a government bailout. After paying the fee, every bank is entitled to government protection on it's CDS portfolio either via sales or insurance.

Thus the non-dealer position fee can force the banks that were engaging in the toxic aspects of the CDS trade to bear the costs of the bailout. Will these banks refuse to participate in the toxic asset purchase/guarantee plan? Possibly, but if so they can be required to mark their books to market and will most likely be nationalized due to insolvency. (If too many banks refuse to play ball, the government can "buy" the CDS from the true dealer banks and then play hardball with the CDS counterparties.)

How is that asset guarantee program s'posed to work?

What makes anyone believe that the troubled banks can afford to pay a reasonable insurance premium on their toxic assets?

For example, the NYT has an article describing one MBS. The MBS is composed of 9,000 second lien mortgages and let's assume that the face value of each averages a little over $50,000 so the total value of the MBS (all tranches) is $500 million. Given the valuation the bank puts on the bond, this may well be the senior tranche, so let's assume the bank has a $250 million senior exposure to losses on these mortgages (i.e. the portfolio needs to lose half its value for the bank to lose a penny). The article states that the bank carries this bond at 97 cents on the dollar or $242.5 million.

However, as anyone who follows the mortgage market knows, second lien loans suffer very badly when housing prices fall dramatically -- because in case of foreclosure there's unlikely to be anything left for the second lien holder. Therefore there is some significant probability that a very large number of these mortgages go into default -- with a recovery of zero.

That's why S&P stated that in their worse case scenario the value of the bank's tranche was worth just 53 cents on the dollar or roughly speaking that only $132.5 million of principal will be recovered (assuming $500 million of loans and a senior tranche of $250 million). Let's say there's a 20% probability that this worse case scenario materializes.

In this case the insurance premium collected by the government insurer of the senior 90% losses of this particular tranche needs to be well over 0.20 ($242.5 - $132.5 - $24.25) = $17.15 million (since a full analysis would not assume that with 80% probability losses are less than $24.25 million and therefore cost the government nothing). In other words for assets of this quality the bank would have to pay significantly more than 7% of the face value of the assets as a premium. Spreading this premium over five years would result in a charge that would have to be well over 1.4% of face value per year.

By contrast in the Bank of America deal that is being used as model for this type of insurance the Fed is charging only 0.2% per year in premiums. Of course, the correct premium depends on what's in the portfolio -- which is not public information -- but it is certainly possible that the Fed is charging only a fraction of a fair market premium for the insurance policy.

If the Fed follows the policy reported in the press of insuring the assets that haven't been written down yet (but like the bond in the NYT article should have been written down), then either the premiums need to be very high relative to the face value of the assets or the government will end up tranferring massive amounts of value to the banks and their shareholders.

In short, if paying banks a fair price for their bad assets is likely to render them insolvent, then charging a fair insurance premium for those assets is also likely to render them insolvent. The only "advantage" of insurance over the purchase of the bad assets is that the government gets to duplicate the outrageously stupid behavior of our bank managers -- they get to make optimistic estimates that indicate that the losses will never show up -- until they do.

Ongoing dialog on derivatives

Apparently the Atlantic Business Channel has a long delay in posting comments, since Derivative Dribble submitted his reply shortly after my post went up, but it didn't appear on the blog for at least 6 hours after that (I checked). Anyhow, since comments are so delayed at the The Atlantic, I thought I would go ahead and post my response to DD here. (These are approximate. I didn't actually copy the posts before hitting submit -- someday I'll learn!)

I'll keep updating this post if the dialog continues.

Derivative Dribble (DD) responded to my post with the following.

ACC,

As I've said a million times, CDSs, like all contracts, create counterparty risks and therefore can create substantial reliance costs. That is not at issue in this article.

Secondly, unfunded derivatives cannot be considered "allocations of capital" since they are, by definition unfunded. That cash is free to circulate through the capital markets.

Posted by: Derivative Dribble | January 29, 2009 4:57 PM



My comment 1: Two questions:

(i) How is counterparty risk different from credit risk? They both involve the risk that a counterparty fails to make payment on a liability.
(ii) Unfunded derivatives are nothing more than a promise to make payment in the future. What “cash” are you talking about?



My comment 2: Just a note for the blog administrator: The delay between the submission of comments and the posting of comments is irritating.

Understanding the Demon in CDS

This is a response to Charles Davi's post at The Atlantic Business Channel. You should read it before reading this post.

Davi writes: I often note that derivatives cannot create net losses in the system. That is, they simply transfer money between two parties. If one party loses X, the other gains X, so the net loss between the two parties is zero.

This argument implicitly assumes government subsidization of “too big to fail" banks. All derivative contracts create counterparty risk, which is just another name for the credit risk associated with the derivative. Because sudden price movements (such as we have seen regularly in financial markets lately) may force the debtor counterparty to default on the CDS (and declare bankruptcy) before enough collateral has been transferred to the creditor counterparty to cover the obligation, one needs to consider what happens in the circumstance that the debtor (or losing) counterparty defaults suddenly: then not only does X contribute to the debtor counterparty’s unpayable obligations, the creditor counterparty does not gain X. Conclusion: not only can “poorly underwritten mortgages can create net losses,” but a poor choice of CDS counterparty can also create net losses – and be evidence of inefficient allocation of capital

The only way to avoid net losses as a potential outcome of trading derivatives is to make it impossible for debtor counterparties to go bankrupt. Effectively that is what the government has done by guaranteeing the debt of the major money center banks. The fact that Davi’s argument can be sustained is not a function of the nature of derivative markets, but rather of State intervention in derivative markets.

I can agree with Davi that the ability “to express a negative view of mortgage default risk” contributed to the popping of the mortgage bubble that was created by the securitization process. However, the fact that popping the bubble took a long time (specifically from January 2006 to August 2007) meant that "too big to fail" financial institutions were able to rack up huge losses by selling CDS insurance (often indirectly through off-balance sheet entities) before the bubble was good and truly popped.

When Davi writes “depending on whether the synthetics are fully funded or not, the principal investment will go to the Treasuries market or back into the capital markets respectively,” he exhibits a fundamental misconception about the nature of CDS that David Harper has pointed out repeatedly on Davi’s Derivative Dribble website.

It is obvious to anyone who compares the size of the CDS market with the size of securities markets in the US (and indeed the world) that “the principal investment” (i.e. the notional value of the CDS) does not get invested in Treasuries or capital markets. When synthetics are not funded, they are backed only by the promise of the protection seller to make good on the contractual obligation – that is, they are pure counterparty (aka credit) risk.

The buyer of protection sold by a partially funded synthetic or hybrid CDO is short the credit risk of the entities referenced in the CDS sold by the CDO and long the credit risk of the super senior counterparty. (The super senior tranche of a synthetic or hybrid CDO is almost always an unfunded CDS.)

Davi makes this claim: “Only when interest rates on MBSs drop low enough, along with the price of protection on MBSs, will protection buyers enter CDS contracts.” This is true only if the efficient markets theorem holds and interest rates instantaneously reflect expectations of default. In general, when expectations of MBS default rise, protections buyers will enter CDS contracts over time driving up the price of protection on CDS and mortgage interest rates. As noted above it took a year and half for this process to play out in financial markets.

The problem with CDS is not intrinsic to the derivative contract itself, but the fact that "too big to fail" banks were allowed to sell protection (when the buyer did not have an insured interest) in the market.