Finance and the Public Interest

Professor McDonald,

I don’t think I understand the CDS graphs here because I can’t reconcile them with bond prices.  If insurance on a Ford bond is around $1800 per $10K, that’s effectively an 18% cost/yr on a bond.  According to Yahoo finance, Ford’s 2016 bond is yielding 9.091%.  Doesn’t that leave a net return of -9%?  Why would anybody hold that bond?  Where is my disconnect?  Thanks,

Quentin

Quentin, this is an excellent question, you’re right that the Yahoo bond data and the CDS data I posted is inconsistent. The answer seems to be that the Yahoo bond price data is wrong. Prof Lucas had previously pointed out this discrepancy and I e-mailed a fixed income trader about this. He sent me Bloomberg screenshots (screens available only to traders, with bids and asks along with names and phone numbers of specific market makers) showing Ford bond yields that were consistent with CDS and a lot higher than those on Yahoo. I don’t know what’s going on with Yahoo.

Prof. McDonald:

What is the reason to believe that CDS data is more reliable than bond data? 

My very uninformed impression is that CDS quotes are not based on actual trades necessarily whereas bond data is. 

In other words, I am trying to understand whether CDS or bond price data is more prone to moving away from fundamentals. 

Any insights would be much appreciated. 

Also, is the data underlying the CDS screenshots in your
post available via a paid (or free) data provider?

Thanks!

Ufuk, this is a fair question. I’m not a trader, but every conversation I’ve had with traders suggests that CDS quotes are real, and generally reflect what is happening with bond prices. One trader told me that during September, bonds were if anything trading at a discount to CDS (the implied CDS quote from the bonds was higher than the actual CDS quote). I can’t quote a measure of liquidity for CDS, but there is certainly a lot of it outstanding, and the prices that are quoted are used by dealers to mark their books to market. (I would be surprised to learn that different institutions are using radically different marks for CDS, but of course the world has been full of surprises the last few months.)

I use Datastream. CDS quotes are also available from Markit and Bloomberg.

One interesting technical aside is that CDS with high premia trade at an upfront fee (points) plus 500 basis points per year. However, the commonly available services quote prices only as an annual premium (they are amortizing the points). This can lead to ridiculous reported premia; Bloomberg a few weeks ago was quoting GMAC CDS at 12,500 bps. I believe this was due to a flawed amortization algorithm. I’m very sorry I didn’t take a screen shot of that one!

Bob

That’s very interesting.  I will look into this further.  Thank you!

Bob,

I am still trying to understand the nature of the CDS markets and their ability to price risk fairly. 

Could you shed light on the following? 

I came accross CDS quotes for debt issued by the U.S. Treasury in the range of $30,000 annually for 5 years.  It looks like this premium went up 4-5 fold in the past year due to the crisis, bailout, etc.   

The question is:  Why would anyone pay anything to buy insurance against US treasury default?  How does the buyer expect to collect on that event should it occur?  If the least risky issuer in the world defaults, who in the world can still be standing to make good on this insurance? 

Further, why would the U.S. Treasury default be of any realistic concern since the issuer has the unique ability to print currency should it run out of it? 

Thanks…

Ufuk

Ufuk,

Yes, CDS on US Treasuries have recently been in the range of 30 basis points. Your points are good ones and I don’t have great answer. Several comments:

1. I asked several traders about this market when I first heard about it and their guess was that these trades are sovereign-to-sovereign (or at least written by a sovereign).

2. The US Govt did apparently almost suffer a techncial default in 1996 when the Republicans refused to approve the budget (Washington shut down for a week or two at that time, don’t remember the details). Even if there were a technical default, however, loss given default would be approximately zero.

3. The quantity is small. The DTCC shows 4 billion notional gross and a little over 1 billion net. It’s not a huge market, but it is a market.

So yes, it’s odd. I don’t have a good explanation.

Bob

Is there any sites on the web which offer CDS prices that arent pay services?

Chris, to my knowledge CDS price quotes are only available via pay services such as Bloomberg, Datastream, and Markit.

Bob

Professor McDonald,

I am glad to have found this post. Can you tell me where I can find live quotes for credit default swaps? I have searched for a while and cannot find anything. I can’t seem to find anything on my brokerage accounts either. Thanks in advance!

You may disregard my previous email/comment as now I have read the comments and you’ve already answered my question. Thanks for the great info! This blog has been bookmarked!

Prof McDonald,
Could you please explain how the CDS Market applies in potential restructuring situations such as GM. In essence, if i am the owner of GM bonds and also have a CDS contract on those bonds am I not indifferent to get $.05 or $.50 for those bonds since the CDS will make up the difference?  Doesn’t the existence of CDS have a material impact on a voluntary restructuring of a company like GM or anyone else going into bankruptcy?

Dear Professor McDonald,

I’m just starting my diploma thesis in which I want to use information about CDS from Datastream. Unfortunately all papers and books I found for my topic use only Bloomberg for CDS data. Therefore I’m writing you this message and hope you could help me.

In my diploma thesis I want to compare two models. The main question is, if one model explains CDS Spreads better than the other one.

I found out, that I could choose in Datastream for the time series different data types but even the description doesn’t tell me exactly what’s the difference between e.g. the data type “Credit Default Swap Premium Bid” (BPB) and the data type “Bid Rate”/“Spread Bid” (SB). If I choose both data types for the same company than I could see that there are little differences between the values but I can’t understand the what’s the reason.
But the main problem is that I couldn’t find out if - for example - the data type “Credit Default Swap Premium Bid” (BPB) is already the spread of the CDS.

Because of your experience with Datastream I send you this message and hope that you could help me.

Kind regards,

Alexandra Roedel

Alan, if restructuring is not a “credit event” (which it seems not to be for existing GM CDS), then a bondholder with a CDS contract has no incentive to restructure. The FT had an article recently about this.

Alexandra, I’m not sure what the difference is. I’ve been using “CDS premium mid”, which seems always to be the average of “CDS premium bid” and “CDS premium offered”.  In the cases I’ve checked, Bloomberg and Datastream give comparable quotes, but I don’t recall the exact differences, if any. With both Datastream and Bloomberg you need to be aware of too-large stated premia when the CDS premia are high. I mentioned this same point above in a response to Ufuk Ince.

The data on credit default swaps in Datastream and Bloomberg comes from CMA (http://www.cmavision.com/).

The following information about the DCS data available through Datastream has been kindly provided by the Thomson Reuter support desk:

CMA receives credit default swap (CDS) prices (“spreads”) from a range of market contributors. These contributors consist of both buy and sell side institutions active in the fixed income markets such as asset managers, hedge funds and banks. These active market participants provide CMA with both real-time and delayed prices of executed trades, firm or indicative bid/offers on a specific entities (e.g. company or emerging market), tenors, seniorities (ranking of the debt receiving moneys in case of default) and restructuring types (definition of what constitutes a default, ISDA agreement types). To ensure the highest level of accuracy CMA checks these prices against previous quotes and validates those using related securities and news.

For less liquid entities where market activity is infrequent, CMA calculates the fair CDS spread using a proprietary issuer/sector curve model that derives an appropriate curve using known liquid CDS spreads, bond spreads and ratings data.

CMA assures full transparency for its clients by providing a qualifier for each point of a specific CDS price time-series. This qualifier indicates the liquidity or if applicable, the extent to which a value has been model-derived.

Market activity among CMA’s sources varies due to their requirements to buy or sell default protection. Bank portfolio managers tend to buy CDS, i.e. they buy protection, to hedge concentration in a specific industry/sector in their loan portfolios, and they sell protection, to subsidize their hedging programs and diversify their loan portfolios. Asset managers are active in relative value trades, to customize the credit risk of their bond portfolios and also use of CDS as substitute for short selling bonds. Hedge funds are active buying and selling protection in accordance with their general investment strategies, e.g. convertible arbitrage, short selling and basis trades (bonds versus CDS, equity versus CDS).

These are the datatypes available for CDS provided to Datastream by CMA:

SM Mid Rate: This shows the mid rate spread between the entity and the relevant benchmark curve. The rate is expressed in basis points. This is the default datatype.

SB Bid Rate: This shows the bid rate spread between the entity and the relevant benchmark curve. The rate is expressed in basis points.

SR Offer Rate: This shows the offer rate spread between the entity and the relevant benchmark curve. The rate is expressed in basis points.

VV Observed/Derived Indicator: This shows the type of price used in calculating the bid and offer rate. The score is expressed numerical as 1 or 2. 1 is the numerical value for an observed market spread and 2 is for a derived spread.

RT Restructuring Type: This shows the type of restructuring for each series. Each type will be expressed in numeric form.
o 1 = MMR (Modified Modified Restructuring): ): This is a “modified” version
      of the modified restructuring option, which resulted from the criticism
      of the modified restructuring that it was too strict with respect to
      deliverable obligations. Under the modified-modified restructuring,
      which is more popular in Europe, deliverable obligations can be
      maturing in up to 60 months after a restructuring.
o 2 = MR (Modified Restructuring):  Modified restructuring has become
      common practice in North America in last few years, which limits
      deliverable obligations to bonds with maturity of less than 30
      months, after a restructuring.
o 3 = CR (Credit Restructuring): This allows the credit event to be triggered
      when a debt obligation is restructured, for example, by having interest
      payments reduced, having its principal amount reduced, having its
      maturity extended, becoming subordinated to another obligation or
      having its currency changed. There is no maturity limitation on the
      kind of obligations that can be delivered following this version of
      restructuring. This is currently used for most emerging market credit
      default swaps as well as for Japanese credits.
o 4 = NR (No Restructuring): This option excludes restructuring altogether
      from the contract, eliminating the possibility that the protection seller
      suffers a “soft” credit event that does not necessarily result in losses
      to the protection buyer.

To compare the risk assessment between two companies of the same sector (i.e. C….S5 CITIGROUP INC SEN 5YR CDS and COF..S5 CAPITAL ONE BANK SEN 5YR CDS ), it would be best to use the SM datatype, for it simply is the average of the Bid rate and the Offer Rate. A rising Mid rate such as that of Citigroup implies that its CDS is more active due to the recent news it has been stirring. You can also do this analysis by cross-sector; i.e. comparing COF..S5 CAPITAL ONE BANK SEN 5YR CDS, F….S5 FORD MOTOR CO SEN 5YR CDS, and CPB..S5 CAMPBELL SOUP CO SEN 5YR CDS. If you’ll be using the 401A program of DS Windows, input all three codes and you will get the graphical report that simply says that the US Motor industry has an erratic trend versus the bank industry, with the food, as represented by CAMPBELL can be viewed as more tamed.

The Bid rate is actually the buy rate quotes as the CDS contract essentially “is an agreement between a protection buyer and a protection seller in which the protection buyer pays the protection seller a fee and in return the protection seller pays the protection buyer the par value of the reference obligation in the event of a default.”

Please remember that below are the cash flows in a CDS transaction:
* If there is no credit event, the only cash flow is the premium payments
  that the protection buyer gives to the protection seller
* If there is a credit event, the protection seller must deliver the notional
  amount of the CDS Contract
* Notional amount—think of life insurance.  You buy a USD 2million life
  insurance, but this is a notional amount.  Why? Your beneficiaries
  receive this in case of your death.  In the meantime, you pay a premium
  equivalent to a certain percentage of this notional amount every year.

Additional information—Credit Events that Trigger CDS Contract:
* Bankruptcy—insolvency of the reference entity; receivership
* Failure to Pay—failure to make payments due to its obligors
* Restructuring - credit event is triggered when a debt obligation
  is restructured which may involve
* reduction of interest
* reduction of principal
* postponement of interest/principal
* subordination of obligation
* change of currency/composition of interest or principal