How TRACE helped kill the dealers

Back in 2002, regulators at the NASD decided 'mom and pop' needed protection from the big bad dealers while they were investing in the corporate bond market.  Under this guise of investor protection, they implemented a program called TRACE (Trade Reporting and Comparison Entry Service).  I use the term 'guise' because individual investors are only a tiny portion of the overall volume of the market.  Most individual investors gain their exposure to the bond market via mutual funds.

This may seem archaic to you equity investors, but until that point, virtually no trades in the corporate bond market were recorded and disseminated broadly.  Yes, up until this point, the dealers and their large institutional clients, had an informational advantage.  Bond trading was long the engine room of dealer profits.  

This opaqueness had it's advantages for the large investors as well.  For instance, if a large holder of XYZ corporate bonds (client 'A') decided they no longer liked the credit, they would choose a dealer to begin 'working' out of them.  The dealer would buy $50mm at a spread of, let's say, 100bps (over similar maturity treasuries).  They would quietly take these bonds on to their balance sheet and begin to try to sell them to other institutional investors.  Suppose client 'B' bought the first $10mm at 90bps (for a profit of 10bps to the dealer).  The dealer would now market them to a new batch of investors (clients C through F).  If the process worked, and it often did, the bonds would be successfully distributed from client 'A' to handful of other investors and the spread (i.e. price) would remain largely intact and the dealers would make money....often a boatload.  

Now, we enter the post TRACE world.  If a block of bonds 'prints' on TRACE, the whole world knows that someone is long a slug of risk and that they bought it at 100bps.  If a dealer calls a client and says "we have a small piece of XYZ to offer", the client is likely to say "no, you have a block of them and I won't pay you any more than 98bps" (remember the dealer bought them at 100bps).  So, the dealers ability to make a profit on the bid/ask is subtantially reduced...especially in light of little to no informational advantage.

In this new post TRACE world, dealers rarely take bonds (i.e. risk) on to their balance sheet.  We've seen a dramatic and wholesale shift from a principal model to an agent-only model. 

In the corporate bond market, you cannot have liquidity without capital.  If the dealers cannot make a consistent profit, they will not allocate capital.  No capital, no liquidity.

TRACE is one of the primary reasons that dealers shifted their capital from the trading of corporate bonds (i.e. 'cash bonds') to the trading of credit default swaps.  CDS trading is still opaque and the informational advantage still lies with the dealer.