Block TRACE again

September 29, 2006

So I checked out the TRACE website the other evening, after following up Accrued‘s post. There are different rules for reporting blocks, obviously designed to protect dealers offering block liquidity.

BIDs and blocks

September 28, 2006

Thanks to Matt for the news on Bids – a new block trading system. Since the system is completely anonymised, I can’t see how it solves the information asymmetry problem for the trader offering block liquidity. How can he be sure that the liquidity taker isn’t better informed, and about to leave him with a massive position against which the market then moves ?

Sparklines

September 26, 2006

Sparklines is a visualisation technique I’ve been following for a while – so I read this with interest. And then something else on cmiles blog caught my eye: a link to Juice Analytics. They have a nice blog on Excel hacks, which mentions a toolkit that gives you sparklines in Excel. Not the same Juice as founded by Charles Ferguson who wrote High Stakes – probably the best book I’ve read on founding a software start up.

TRACE the blocks

September 26, 2006

Accrued Interest posts on a study claiming that the TRACE corporate bond trade reporting systems had collapsed spreads to the tune of $1 billion. There are many other factors that could affect spreads beyond TRACE, so I question the attribution of all of that billion to the reporting effect. Accrued points to increased competition – a very plausible explanation as the study period coincides with the rise of electronic trading for bonds, which has surely driven down spreads.

Reporting could cause spreads to widen for block trades – especially for the less liquid securities. If an institutional trader wants to sell a block of an illiquid corporate bond, then any dealer offering liquidity will have to price in the cost of front runners taking the trades he’ll need to make to unwind his inventory when they see the block reported. Any front running trader who’s long will see the block reported, sell his long position, and buy it back cheaper when the dealer drives prices down selling off  his block.

Harris on block trading

September 25, 2006

Why do we tend to see larger deal sizes on dealer to client ECNs like TradeWeb than we do on interdealer ECNs like MTS ?  As usual Harris is illuminating in chapter 15, on block traders.

When a dealer is approached by a trader demanding liquidity in large size, he will suspect that trader of being better informed. After all, if you’re a well informed trader about to place a winning trade, wouldn’t you size up the deal to maximise winnings ? The dealer will worry that if he takes the trade, prices will move against him, and he’ll lose money on a position he can’t unwind. He’ll also worry that the block may have been split up, and there may be more large trades to follow. So any dealer taking a large trade must be sure that his counterparty is less well informed. This isn’t a problem on quote driven markets, where we know the counterparty identity when handling an RFQ. But that can’t be the case on a market organised as an anonymous order driven exchange…

FPGAs & autoquoting

September 23, 2006

NewYorkScot summarises the High Performance on Wall St conference. Nice one! I guess we won’t be putting our pricing engine on FPGA any time soon. While minimal latency pricing is critical for fixed income autoquoting systems, flexibility is top priority too. Those crazy traders want to twiddle their pricing models intraday. And they want new pricing models all the time too…

Matlab, Python, R

September 22, 2006

Thanks to Vince for this Matlab/Python/R crib sheet. It’s good to see the three way equivalence, especially since I’m an old hand with Python and an R newbie. Matlab is the tool of choice on our trading floor for traders who’ve hit the buffers with Excel.

Still learning with R

September 20, 2006

In all my R coding so far I’ve been referencing individual lists within a dataframe (trans: column in a table) with the frame$list syntax. Not very generic, especially when I’m trying to write a generic, multi dimensional bucketing function. So somewhat belatedly I’ve discovered the frame[[list]] syntax, which allows me to parameterise the column references I pass to my bucketing code. Phew!

Insider trading

September 19, 2006

Just had another one of those Harris moments reading 14.6.2.1 in Harris on spreads. He points out that “markets that effectively enforce insider trading rules protect their liquidity suppliers from adverse selection.” Since adverse selection widens spreads, this benefits all traders relying on dealer supplied liquidity to execute their investment strategies.

Spread determinants

September 17, 2006

I posted on Harris’ discussion of spreads earlier. Harris goes on to summarise the main three spread determinants as asymmetric information, volatility and utilitarian trading interest. Utilitarian interest is trading by investors, hedgers and gamblers rather than market makers. Volatility can be measured, but the other two factors can’t.

When there is a high degree of asymmetry in trader information dealers will be hit by the adverse selection of the well informed traders. Consequently they’ll set spreads wide to recover the costs of adverse selection from uninformed traders.

Which makes me wonder how asymmetrical information can be in the fixed income rates world. I can imagine all kinds of asymmetry in the equity world, from insider knowledge to detailed sector research. But what is there to know about a straight government bond ?  It’s principal, coupons and maturity. Given those we can calculate its present value. Of course, it’s not quite as simple as that, but my point is that there’s no potentially hidden information. This should lead to narrow spreads.