Actionable insights on equities, fixed-income, macros and personal finance Start 14-Days Free Trial
Actionable investing insights Get Free Trial

On Yahoo: Credit Default Swaps

I write on Yahoo about Credit Default Swaps:

A week of fantastic weather and gridlocked traffic in Bangalore makes the old garden-city-dweller in me long for the Bangalore of a distant past: how the last 10 years have changed it from a lazy retiree town to an info-tech powerhouse, and in the process introduced it to multi-hour traffic snarls and high levels of pollution. While growth has had its good points, the city’s infrastructure just hasn’t been able to keep up.

The downside of superfast growth was reflected in a book I read: “Fool’s Gold“, by Gillian Tett on the way greed worked itself through the credit derivatives and banking market in the US. The history of the “credit default swap” – the derivative most often quoted in credit-derivative land – involved Exxon and J.P. Morgan Bank during the Exxon Oil Spill (mid-90s). Exxon wanted a $4.8 billion loan from JPM, and was facing a potentially massive damage ruling against it because of the spill. Ordinarily, a JP Morgan banker would hear such a request, excuse himself from the room, laugh his head off in the corridor, return and say “No”.

But you don’t do this kind of thing with a big client like Exxon; you risk losing their future business, and competition will make inroads. But the risk of the loan was still too high. So, the good folks at JP Morgan decided to lend Exxon the money, but then pass the risk of the loan over to the European bank of Reconstruction and Development. Meaning: If Exxon defaulted in paying J.P. Morgan back, the E.B.R.D. would make it up. A “credit default” meant that the liability was “swapped” to someone else. E.B.R.D. would be paid a sum of money that made it worth their while, as a premium. Thus, a Credit Default Swap was born.

When the Exxon default risk was taken out, J.P.Morgan now did not have to reserve money against the loan – a regulatory requirement for any loans that involve a risk of default – so they could lend their money to more creditworthy borrowers, and still keep Exxon happy.

The Credit Default Swap (CDS) market then started to grow, as banks realized they could optimize capital reserves if they hived off the risk to someone else. They lent, they bought CDS protection and that freed tied-up reserves to do even more lending. As an example: a bank that might have $100 of capital and could lend, say, upto $1,000 using a 10% reserve system, could buy credit default swaps for $600 worth of loans, and then lend that $500 to someone else. They now ended up with $1,500 worth of loans, $500 worth CDS (for which they have paid some premium) on the same capital base of $100. This was “okay” under regulation, and in principle would have helped the system by diversifying risk.

As the market scaled, CDS prices became more important. Where markets would price default risk in falling prices of bonds (for example, a Rs. 100 bond trading at Rs. 70 might mean people expect to be paid only 70% of the loan back) such default assumptions might be reflected in increasing CDS prices. A “proxy” for default risk, and a tool to diversify – a win-win proposition?

Not entirely true. You can buy or sell CDS without owning the underlying debt, in effect speculating on a default or the lack of it. If there’s no default, a buyer will lose some premium – of the order of2-3% of the total amount he has bought CDS for. A default may give him, say, 30 to 50 per cent of the amount insured. So I could buy a CDS on $10 million worth of a company’s debt, by paying $300,000 per year – and if the company defaults, I might get back $3 million.

A 10x return on investment is the kind of leveraged bet that attracts traders – and in effect, a lot of people buying CDS will push up the prices, and therefore give the “appearance” of a higher risk of default. In a world increasingly focused on posturing rather than action, this is considered bad – even European politicians decried “naked CDS” traders as being responsible for worsening the Greek crisis. But what it really means is that CDS prices aren’t very good indicators of actual default, or that the worsening CDS prices might have prompted anti-default action that would otherwise have not materialized.

The other issue with credit default swaps was the underlying debt itself. A company or country default is a simple thing – if the entity fails to pay back principal or interest, it’s considered a default. But this wasn’t satisfying enough – after a while, everyone wants to sell insurance on the good companies and buy it on the bad ones. Bankers then decided to move the concept to “securitized” loans – where they would pool the loans of multiple companies together and sell bits and pieces of the merged loans as securities to other entities like pension funds.

A straight securitization would mean interest would flow equally to all holders, and any default would hit everyone equal. But that wouldn’t fly with investors – some wanted higher returns for higher risk, and most just wanted a good return for very low risk. Financial innovation came in again: Enter the Collateralized Debt Obligation (CDO). The pool was sliced into “tranches” – designed such that the defaults would first hit the lowest tranches, and only then flow upwards. After all, with a large number of loans you may not ever see very high default rates, so the higher tranches are safe (just like you don’t usually see all the people in a city have an accident at the same time, unless you are in Bangalore where it’s very likely) In a pool of 1,000 loans worth $100 million, you might expect only 15% defaults – that means the first 15% was “risky” and the tranches above them lesser so. In return, the highest tranches got the lowest interest rates. These tranches were rated by the rating agencies, whose inputs were primarily computer models because there was really nothing else available.

Credit Default Swaps were used to offset the risk of most such tranches. The highest (“super-senior”) tranches would have such a low CDS premium that banks preferred to retain them on their books (though a number were sold to AIG). Bank books were now bloated
The market took off substantially, not just in the corporate debt area, but in the mortgage debt space (where the pools were made of housing loans). Banks did not just offset risk; there was so much demand for such instruments that they could charge hefty fees when they created CDOs.

Growth became so huge that when it wasn’t possible to lend more, bankers would take the risky tranches – that no one wanted to buy direct – and mix them in another CDO, and tell people that the highest tranches of THOSE CDOs were less risk. Using complex mathematical formulas and with limited default history, this concept also sold.

And all of this was in a period of low interest rates, where pension funds desperately needed to make more than the standard non-risky avenues offered.

The consequences: When you can give loans and “sell” the risk away through such instruments, you don’t really have to bother about the actual risk of the loan – something the banks were supposed to be doing. As a result, horribly bad loans were originated and sold away. The rating agencies were earning substantial fees from rating the higher tranches of the CDOs, and therefore reacted as late as possible to downgrade when defaults increased. Banks were supposed to have been beneficiaries; instead, they tanked up on the CDOs and CDS liabilities themselves, leaving themselves exposed to the massive increase in defaults when it actually happened. Regulators refused to curtail these markets and banks grew so big that their failure would threaten the world economy.

When defaults grew, those who sold CDS contracts suddenly found themselves in a spot – their complex computer models told them this should not happen, yet this was happening. A number of them defaulted, yet others like AIG were taken over by the government in a rescue. (It’s a different matter that those who got rescued were top banks that had bought CDS contracts from AIG) It eventually ballooned into a worldwide financial crisis, and we are still seeing the repercussions.

Credit default swaps are blamed for worsening the crisis; even legendary trader George Soros has at asked to outlaw them or at least, make them only available to people owning the underlying debt. The “must-own-bond” concept does hold some water; any unbridled speculation should only be possible on regulated exchanges with strict leverage and position limits. The industry’s working hard in both directions – some to create and others to stymie regulation in their otherwise unfettered world. Let’s see who wins.


Oh, and here’s how to get the book:



  • Anonymous says:

    >What is the difference in arguments against naked shorting/puts vs naked CDS?

  • Deepak Shenoy says:

    >Naked shorting (not possible in India, only in the US where you short without borrowing first) is a stupid concept.

    Shorting futures is not "naked" – there's a buyer on the other side.

    Futures and Puts (options) are regulated through exchanges (there is centralized clearing, collateral management and position limits) which are again fine. CDS has had major issues in all these areas, plus the counterparty risk.

  • Anonymous says:

    >I was wondering in context to the "must-own-bond" concept.

    Why should buying a CDS require owning the bond as opposed to say buying a put (which does not require owning the stock). Assume both transactions are done with an exchange (to ignore the issues you pointed out).

    The argument made is one of perverse incentives. But, the same exists when owning a put – it is a speculation that the company does badly. So my question is, is there anything different between these two cases?

  • Deepak Shenoy says:

    >Anon: The must-own-bond concept is to make sure there is insurable interest, to work for really removing the risk on bank books. One example: if I buy two CDS and I own one bond, the risk being written off cannot be 2x (that skews the metric, because the other CDS may have counterparty risk, may not be on the same sort of bond, differing underlying securities or terms etc.) Unfortunately that is allowed to happen today.

    While a CDS is a put, and you should be able to bet on puts, the concept has gotten so big that it introduced systemic risk due to the non-transparency, non standardization, counterparty changes (don't know who's on the other side), different default triggers etc. So it's best this kind of speculation goes on an exchange where there's central clearing, and regulation on limits.

  • Deepak Shenoy says:

    >Anon: I didn't mean to mention perverse incentives in the context that people should not buy CDS to speculate. Speculation is good, it just changes the purpose – the reason you can't gauge default risk from CDS prices is that there is enough speculation to skew that. But if you look at CDS, it's a good tool to reduce risk – it's market price can be an entirely different thing.

  • Anonymous says:

    >I didn't understand your point about the metric being skewed – 2x risk being written off. However, you seem to be suggesting that the "must-own-bond" concept should apply atleast to financial companies backed by the taxpayer. ie; it comes under the umbrella of.. banks should not trade on a proprietary basis (which is the Volcker rule). They probably shouldn't be making big bets via puts, commodities futures, etc and so the same should apply for CDS? Hedge funds on the other hand should be able to speculate freely.

    You make the point that such speculation is best done on the exchanges. No doubt, and I explicitly assumed that that was the case for the sake of argument. Your key issue is counterparty risk which is orthogonal to the must-own-bond issue. I was interested in just the must-own-bond concept.

    I don't agree with your second comment. Speculation affecting/skewing market prices is a long studied issue which encompasses every type of asset class. You could make the same argument with commodities, for example – ie; Indian govt banning commodity future trading because of the allegation of speculation skewing prices. I'm not sure that CDS has a worse skewing effect than say puts or futures on the underlying asset.

  • rajsmusings says:

    >I missed the story line a li'l bit on this para

    "As an example: a bank that might have $100 of capital and could lend, say, upto $1,000 using a 10% reserve system, could buy credit default swaps for $600 worth of loans, and then lend that $500 to someone else. They now ended up with $1,500 worth of loans, $500 worth CDS (for which they have paid some premium) on the same capital base of $100."
    Is it possible to explain the example a little more ? Or may be i need to increase my financial literacy more 🙂

    Above all i enjoy reading your post's very much. Keep the good work!


  • Deepak Shenoy says:

    >Anon: About the 2x leverage. A simple example is if a bank has $100 worth of a certain kind of loan (but another $1000 of another kind) and buys CDS for $200 worth of it. Typically you would offset the $100 risk and assume the other $100 is naked open. But the way bank risk metrics work is that they assume all sorts of things and offset risk of even the "other kind" $1000 loans because that's how the risk metrics work. If you trade them on an exchange the contracts are standard and you only offset risk of what is really offset-able.

    On the must-own-bond concept: Buying CDS is probably less risky (you can't lose moer than your premium). But most players buy and sell, and offset the risks together. That's why it holds water.

    Must-hold-bonds has the other reasoning that nowadays there is far more CDS on certain types of bonds than the face value of the bonds themselves, which doesn't quite make sense.

    In terms of my second point – I think we're talking at cross purposes. I know speculation skews asset prices. That doesn't mean you can stop it – because the lack of speculation also skews asset prices where there is no liquidity (There are more instances of non-market commodities being manipulated than of market stuff, even as %ages) But that's again besides the point. My point is CDS – or puts or futures – don't indicate the default potential of anything. Speculation doesn't mean the pricing is more efficient at dictating underlying solvency or the lack of it thereof.

    (Though nowadays they might. For financial institutions, where fractional reserves mean that they have to raise capital when there is a run, and that's exactly when the run makes capital more expensive. Isn't quite so for people like BP who have assets that can be sold.)

  • Anonymous says:

    >I don't quite concur why CDS would not indicate appropriate default probabilities. And even if it did not, it would be along the lines of how futures are not good estimates for expected oil prices, etc.

    The fact that total CDS notional may be greater than bond notional might be the only reason for the "must-own-bond" concept. Although, even there one must ask, so what?

    And, who's to say that Greek default was not a real possibility (and still isn't) given their track record? I'd bet that CDS prices are at least better indicators than ratings by S&P.

  • Anonymous says:

    >"a Rs. 100 bond trading at Rs. 70 might mean people expect to be paid only 70% of the loan back"

    You really need to brush up on your bond fundamentals.

  • Deepak Shenoy says:

    >Anon1: Futures prices are not good estimates for expected oil prices, that's a fact 🙂 (Check even last 5 years of data)

    Greek default's probability shouldn't be inferred from the CDS price, where so many other factors can take precedence. They have of course defaulted earlier.

    ANon2: Agree, oversimplification on my part. That's why I had to go with "might". For the curious, a "perfectly safe" bond of Rs. 100 could trade at Rs. 70 because of higher market yields (a 7% bond in a market that where you can get 10% on other ultra-safe for instance), or because of general market liquidity issues like a benchmark change, lack of buyers etc.

  • arghya says:

    >CDS should only be available to people owning the underlying debt – I think it doesn’t hold any water. It is better to sell off the underlying itself instead of buying protection in the name of CDS premium. When you pay the CDS premium in an effect your yield on the underlying got decreased by significant margin. Then why you would hold the underlying at all? It’s a crappy concept like bunch of other thousands.

    The merit of CDS is that bank can offload the RISK(only) to create opportunity for more loans and those who are looking for more return can capitalize this risk. It is fool’s gold for those who are buying the risk (CDS seller) without rationale. And for this you can’t blame the innovation of CDS, blame should lies upon those greedy who want to make quick buck without any rationale of risk adjusted return.

    Hedge funds do not follow naked shorts. A very delicate risk management system has to be in place to deal with the CDS leverages. CDS are excellent risk management tool. A rigorous sensitivity analysis determines the exact amount of long-short requirements of CDS; Overall it may appear that the entire fund is 3x(or more) short but in reality it’s not. Yes of course those who give priority to greed over risk managements are to be on the side of the fool’s gold.