# GUEST COMMENT: The best bond traders don't always accept what they're given

Bond traders are overwhelmed with data, as anyone having access to a Bloomberg terminal knows. The challenge is to identify the numbers that are useful to make a buy, hold, or sell decision. I define information to be the subset of data that has utility in decision-making. The task then is to distinguish between what are useless data and what truly is information. This is an example to demonstrate the problem facing a bond trader using data presented on Bloomberg's Yield Analysis (YA) page.

Consider a deeply discounted MBIA security, the 4.650% bond maturing on July 1, 2018 (CUSIP 55266LAX). On September 21, 2011, the B-rated bond was priced at 71.125 for settlement on September 26th to yield 10.766% on a street convention basis, meaning that bank holidays and weekends are neglected in the assumed timing of payments. This result is easily confirmed using Excel: YIELD ("9/26/2011", "7/01/2018", 0.04650, 71.125, 100, 2, 0). The inputs are the settlement date, maturity date, annual coupon rate, flat price, redemption amount, payment frequency, and the code for the 30/360 day-count convention.

Also reported on the YA page is the projected after-tax rate of return of 8.170%, assuming an ordinary income tax rate of 35% and a capital gains rate of 15%. This too can be confirmed on Excel: YIELD ("9/26/2011", "7/01/2018", 0.04650 * (1 - 0.35), 71.125, 100 - (100 - 71.125) * 0.15, 2, 0). Note that the third entry is the after-tax coupon rate using the ordinary tax rate; the fifth item is the after-tax redemption amount using the capital gains tax rate.

Are these before-tax and after-tax yields data or information? I contend that the street convention yield is information, albeit with some caveats. It measures a buy-and-hold investor's rate of return under the very strong assumptions of no default and reinvestment of cash flows at that same yield of 10.766% (in securities that also do not default). However, the after-tax yield of 8.170% is definitely not information. In fact, it's not even good data.

Under current U.S. tax law, the "gain" from buying the bond at a "market discount" of 71.125 and redeeming the bond at 100 is not taxed at the capital gains rate. Instead, it is taxed as deferred interest income at the ordinary income rate. Entering 35% in the fifth item of the Excel YIELD function gives an after-tax rate of 7.331%--that's a considerable difference.

Who would anticipate that the projected after-tax rate of return on a bond issued by a U.S. company, denominated in U.S. dollars, and reported by a U.S. financial data service would not be based on U.S. tax law? Bloomberg's rationale is that this particular MBIA bond--which is a private placement via its global funding group and registered as a 144A issue, meaning the security initially was sold to qualified institutional buyers--would have been marketed primarily to international investors. Bloomberg claims to use U.S. tax law only on designated U.S. domestic bonds. The problem is that there is nothing on the Bloomberg YA page (or on the Security Description DES page or in help) to indicate this fact or the countries for which this tax treatment applies.

The moral of the story is that the best traders confirm calculations and verify assumptions when using data emanating from "black box" technologies.

Donald J. Smith is an associate professor at the Boston University School of Management and the author of Bond Math: The Theory Behind the Formulas (Wiley Finance, 2011).