Skip to Content

Treasury traders will trade with themselves, when no one else will

US Secret Service Uniform Division officUS Secret Service Uniform Division offic

The Treasury market, officially, had its own flash crash.

On October 15, the yield on Treasury bonds dropped 8% in roughly six minutes. They quickly rebounded, erasing all the gains that bond traders quickly achieved. (Remember, yields move in the opposite direction of bond prices.)

Moves like that don’t happen very often in the low-volatility Treasury market. So, since then, bond market participants have been scratching their heads trying to figure out what happened. And a lot of people have used that unusual day in the bond market to complain about things they don’t like. For Wall Streeters, that has been the Volcker Rule, and liquidity and capital rules, which they say makes it impossible for the big banks to commit a lot of resources to helping their clients trade bonds.

On Monday, nearly nine months later, several regulators published a deep dive on what happened on October 15. Shortly after the bond flash crash, I wrote that there was little evidence that regulators and financial reforms put in place after the financial crisis had broken the bond market. Monday’s report backs that up. Although, just to reiterate, the report was written by regulators, so there probably was some bias there. The report says, “By many metrics, the liquidity and efficiency of trading in the Treasury market are as robust as they have ever been.”

Nevertheless, the report also says that much of the liquidity in the Treasury market has switched from the banks and brokers to “principal trading firms,” which sounds a lot like high frequency trading firms but just by another, less controversial, name.

The report did find that there was an unusual amount of “self-trading” going on during the period in which Treasury yields rose and then fell quickly. Self-trading is when one firm sells a bunch of Treasury bonds to itself. In the stock market, where high frequency trading is much more controversial, it’s wash-trading and is generally illegal. Buying and selling a stock to yourself at ever higher prices is a pretty good way to make a stock price look like it is going up and thus manipulate the market.

But in the Treasury market, self-trading happens all the time and, at least to some extent, is, oddly, allowed. The firms are supposed to have some sort of software, or mechanism, in place, to prevent self-trading, but it appears those mechanisms are optional. Although, last year, high frequency trading firm Tower Research Capital was fined $65,000 for not implementing measures that would limit self-trading.

Despite discouragements, self-trading represents about 5.6% of the Treasury market, according to Monday’s report. But during the bond market flash crash, it reached nearly 15%. That sounds bad.

But Monday’s report makes no judgement on whether all of this self-trading constituted market manipulation or illegal activity. And there is a little bit of a chicken and egg problem. There was a jump in the number of trades placed by high frequency trading firms on October 15. And there was a jump in self-trading during the period in which Treasury yields dove and rebounded. And the high frequency trading firms were the source of the bulk of the self-trades. But it’s unclear whether self-trading moved the market, or if the fact that more of the market was driven by high frequency trading created more self-trading.

What appears to have happened is that, sometime on the morning of October 15, a bunch of computer systems at trading firms anticipated a spike in trading and started to ramp up their activity, as any good market maker should, anticipating the need for higher liquidity. The problem was no actual investors were looking to trade. Monday’s report stated that long-term traders didn’t appear to shift their positions much. When no one else showed up, that didn’t stop the computers. They just traded with themselves, which is sort of allowed. Until they realized this was a little silly and stopped, and Treasury prices went back to normal.

That left the rest of us wondering what the heck happened for nine months. Perhaps. The report does not definitively say that high frequency trading caused the Treasury market flash crash, but it says it’s something to look into.

The Treasury market is more volatile than it used to be. And that’s not great. But if it’s because high frequency trading firms have taken over doing the silly, and riskier, trades banks once did, a higher volatility Treasury market may not be such a bad thing.