Monday, July 06, 2009

Can't See the Forest For the Trees

I can't believe I go distracted from the simple truth behind the NYSE's change in the reporting of the weekly program trading statistics that I missed the obvious explanation. I spent several posts debunking conspiracy theories and blatantly incorrect assessments of the future of the program trading data, but I failed to find the real story, until it jumped out at me this morning.

Let's check the facts again - last week the NYSE announced a change in their reporting requirements for the weekly Program Trading statistics which they report. Instead of relying on firms to submit their volumes to the Exchange, the Exchange is going to take the order entry data right out of the order that is sent to the Exchange, and compile the report itself. The story was totally bastardized into conspiracy theories of "Goldman is trying to hide that its trading all its shares for its own account," and "So much for transparency, THEY don't want us to see what's really going on" when in reality, the new methodology will just make the data more accurate.

I spent so much time trying to debunk falsehoods that I missed the simplest explanation: the NYSE has been losing market share to other marketplaces. In an effort to recapture volume, they are telling the member firms, essentially, that only shares executed on the NYSE will be counted in the weekly program trading statistics. Thus, if you want to draw any sort of foil-hat conclusion from the change, it's that the NYSE is exerting pressure on the firms to trade on the NYSE itself, in order to have their volumes advertised on the report. This is almost certainly true, but I don't see a conspiracy. If the firms want the data advertised, they will trade with the NYSE. If they don't care, they will trade elsewhere.

I am almost certain that we'll soon find out that the mysterious disappearance of former leader Goldman Sachs from last week's program trading stats was a data error. However, if you want a conspiracy theory, forget the one about the Russian spy who stole Goldman's quant trading code which resulted in Goldman working to suppress the dissemination of their trading data in order to hide the magnitude of the theft. A much more logical theory, although one which I still don't think explains what really happened, is that Goldman didn't like the NYSE's "blackmail" attempt to force volume onto the NYSE, and decided to take all of their volume off-Exchange. Since the NYSE didn't announce the reporting change until 6/24, and the most recent program trading statistics are for the week of 6/22-6/26, it seems logical to anyone who bothers to do the work that even this explanation doesn't hold water. Although Goldman almost certainly would have been aware that the NYSE was considering changing their methodology, it's monstrously unlikely that they'd have decided to pull all of their volume during the week of the Russell Rebalance, where they'd need to execute at the NYSE closing price via the exchange.

The NYSE's change in reporting methodology for program trading statistics is an effort for the exchange to recapture volume that has been migrating to other market centers. Firms want to show up in this volume report so that they can market themselves as large providers of liquidity to their clients. If anything, Goldman wouldn't be in cahoots with the exchange to suppress their data - they'd be trying to amplify it. This is a great time for Occam's Razor - the simplest explanation is usually the correct one. In this case, the simplest explanation for GS's descent from number 1 to non-top-15 in the statistics is that either 1) they were trying to prove a point to the NYSE (which seems impossible, as I explained above) or 2) there was a data error, which we'll hear about shortly.


p.s. - the NYSE will still publish stats derived from the firm-submitted daily program trading reports for a few more weeks. After that point, I expect the volume numbers on the report will drop substantially, as all off-Exhange volume will no longer be counted. Of course, when this happens, hypesters will write about the amazing red flag data that is coming out, and I'll have to link back to this post to explain how it was something that could have been easily foreseen by anyone who understands the situation.


Anonymous said...


Thanks for providing some balance concerning zerohedge.

While there are interesting things posted there, the extreme conspiracy thinking is a waste of energy.

Anonymous said...

Why are you trolling Zerohedge? As far as I can tell they don't even know you are alive.

Kid Dynamite said...

i'm sure they don't. my goal isn't to make sure ZH knows i'm alive - I'm just trying to set the record straight. that's all. It's not a war vs ZH either - as I wrote in a previous post, I think he does great work. The problem is, when you write stuff that's blatantly wrong, it casts doubt on all your other work - and in addition, it highlights that you have no idea what you're talking about.

Kid Dynamite said...

note: the truth is out: it's a data error, as expected.

of course, this story isn't quite as sexy and controversy inducing as a russian spy stealing the code of the computer trading program that runs the world, and thus threatening national security.

what happened is either 1) GS traded so much volume during Russell that they overflowed some report cell and got cut off, or 2) the guy putting together the report a day early on Thursday, July 2nd ahead of the Friday holiday was in a hurry and messed up his copy & paste. or both...