Research & Publications
Back-to-Basics

The following article is reprinted from the May, 1997 issue of On the Edge,
the CMS bimonthly newsletter.

Back-to-Basics: Bond Price Transparency?

Teri Geske
Senior Vice President, Product Development


The subject of bond pricing has received a good deal of publicity lately. SEC Chairman Arthur Levitt has called for increased transparency of corporate bond prices, saying, "the sad truth is thatinvestors in the corporate bond market do not enjoy the same access to information as a car buyer or a homebuyer or even a fruit buyer." The Bond Market Association (BMA) acknowledged Chairman Levitt's call to action, stating "the Association fully supports the goal of providing investors with meaningful price information, and we would like to reaffirm our commitment to improving price transparency in the corporate bond markets."

In the BMA's testimony before a House subcommittee this September, CMS was cited as one of the sources of corporate bond valuations. In its testimony, the BMA noted that "there are literally millions of individual corporate debt securities, the overwhelming majority of which trade very rarely," and observed that "the bond markets are so broad and diverse that knowing the price at which one bond traded may say very little about where another might trade."

The BMA is currently soliciting responses to a Request for Proposal for a system that would collect "live" data on corporate bond transactions from inter-dealer brokers for distribution to regulators and other market participants. CMS is closely monitoring this effort, focusing on how this information could ultimately be used to supplement the corporate bond prices in the CMS database. Given these recent developments, we thought it would be useful to review our methodologies for pricing various security types in BondEdge.

The prices in the CMS database are fair market valuations that are calculated and made available daily. Actual traded prices are provided for Treasuries and Futures. We use dealer prices as the basis for computing fair valuations on the over one million securities in our database as follows:

Corporate Bonds and Asset-Backed Securities:We obtain prices from various dealers for a representative sample of corporate bonds and ABS with various maturities/average lives across all sectors and quality ratings. Using the CMS option model, we derive option-adjusted spreads from these traded prices. We use these OAS's to construct a pricing matrix based on sector, quality and duration. We use this matrix to "OAS price" the universe of corporate bonds in our database, incorporating the specific option features of each security. Note that this is more robust than a "nominal spread" approach, i.e. computing a bond's yield and price based on a spread to a single average life-matched Treasury yield.

Fixed Rate Mortgage Pass-Throughs: We start with TBA prices, which are the best indication of where the mortgage market is currently trading. Using these prices, we derive OAS's for pass-throughs across various WACs/WAMs. To price the generic and pool-specific mortgage pass-through in the CMS database, we use the OAS of the TBA whose WAC and WAM is closest to the pool in question as an input to the Representative Path̉ Monte Carlo analytical process. This process derives a price based on the prepayments modeled over the range of interest rate paths sampled, discounted at the Treasury spot rates plus the TBA-based OAS.

  • Adjustable Rate Mortgage Pools: As with fixed rate mortgages, we use TBA prices as the basis for pricing ARMs. We then compute the Bond-Equivalent Effective Margin (BEEM) implied by the TBA prices for different ARM types (e.g., GNMAs, conventional 1 yr. CMTs, etc.). We use this approach because the ARM market does not trade on an OAS basis. When you specify "CMS Pricing" for an ARM pool, we add the appropriate Effective Margin to the underlying index value to compute a yield-to-maturity, which is then used to calculate the ARM's price.
  • CMOs: We obtain prices from several dealers on a large number of CMOs and derive the nominal spread to an average life-matched Treasury for each security. Using these computed spreads, we construct a CMO pricing matrix based on tranche type and average life. For example, 2-year PACs might be trading at 60bps over Treasuries, 5 year Sequentials trading at 90bps over, and so on. We use this matrix to price the universe of CMOs in the database since an OAS-based approach is difficult due to technical limitations. The CMS database contains approximately 40,000 CMO tranches and if we assume that it takes an average of 45 seconds per tranche to run a Monte Carlo analysis, it would take 500 hours (almost 21 days, running 24 hours a day!) to calculate OAS-based prices for all of the CMOs in the database.

Computing valuations on all of these securities requires not only the proper analytical models including term structure, option, prepayment, and reverse-engineering models, but also the data sources (dealer prices, bond descriptive data, updated factors, etc.) and the technology to put it all together. We continually search for ways to improve our pricing and security coverage and will keep you informed of our efforts in this area.

As always, if you have any suggestions for a future "Back-to-Basics" column, we want to hear from you. Please contact marketing at (310) 479-9715 or via e-mail at fia.marketing@interactivedata.com.