Monday, September 25, 2006

Bond Traders Lose $1 Billion?

In a classic example of a headline that means less than it sounds, this from Bloomberg News: Bond Traders Lost $1 Billion From Trace, Study Shows.

The study was about the TRACE system, which requires dealers to disclose trading levels on corporate bonds within 15 minutes of the trade. The idea is that investors can see where bonds have been trading, and are therefore not ripped off by the brokerage firm. Bonds trade with implied commissions, meaning that investors rarely know how much the brokerage firm made on the trade. For example, at any given moment, Merrill Lynch may be making a market in Time Warner 2012 bonds at +110/106, meaning that they would buy bonds at a spread of +110 and sell them at +106. The 4bps differential is the commission to Merrill Lynch traders/salespeople.

The TRACE system has been operating since July 2002, and has been expanded to municipals recently. The study claims that the average corporate bond bid/ask spread narrowed from 16bps to 8bps in the year after TRACE was introduced. The story doesn't make it totally clear how they got from 8bps in tightening to $1 billion lost. It seems like they figured how much tightening occurred on average and multiplied that by how much trading occurred. The study is due to be published in the Journal of Financial Economics within the next 6-months, and was written by Kuman Venkateraman, Hendrik Bessembinder, and William Maxwell.

I have to actually read this study before I believe it. I'm highly skeptical of the $1 billion figure, because as the story mentions, trading overall has increased since 2002. In fact, all I keep hearing about is how much money big brokerage firms are making on bond trading. I also think natural competition, particularly for hedge fund business, is pushing spreads tighter. So some of the bid/ask tightening might be TRACE-related, but some might not be.

I've always thought TRACE wasn't all it was cracked up to be, because many bonds don't trade every day, and sometimes trades are misreported (intentionally or not). So it isn't always good information to look at the TRACE feed for a bond. I think more information is better than less, and I think the Bond Market Association's objections to TRACE are self-serving. So I'm all for TRACE, I just think a lot of retail investors won't know how to read the TRACE feed properly and therefore won't be any better off than before the system was in place.

To be sure, tighter bid/ask is making the retail and mid-sized institutional market tougher. The story mentions brokerages eliminating some/all of their corporate bonds research departments. I hear anecdotally from salespeople that used to do a lot of corporate bond business that times are tough. I know several people who have left the bond sales business. Over the last 30 years or so, we've seen the American consumer choose lower prices over better service time and time again see the rise of Wal-Mart and Southwest Airlines, the decline of full-service gas stations, etc. Looks like corporate bonds are no different.

, ,