Wednesday, July 01, 2009

Two Sides to Every Story

It really shouldn't be this hard. I don't want to spend every day correcting the ignorant rantings of other bloggers, but when they completely misunderstand simple press releases in my area of expertise, I can't help myself.

Last week I addressed some misunderstandings regarding the risk (or lack thereof) in the ETF-Underlying trade. Yesterday, I mentioned how ZeroHedge erroneously got all wound up about an announcement from the NYSE regarding a change in the reporting of the weekly program trading statistics. ZeroHedge received a reply from the NYSE relatively quickly, which they published.

Today, Karl Denninger has a ludicrous rant on his website which gets the story completely wrong, incites his readers who follow him blindly, and stokes more conspiracy theories which are a waste of time. Again - I enjoy reading Denninger - but he and his readers need to understand that not everything is a conspiracy. By "crying wolf" when there is no wolf, Denninger discredits himself, and also distracts himself from his legitimate arguments and causes.

Denninger writes
"what this means, in short, is that the ability of people (like you and I) to see the fact that a handful of banks, most specifically Goldman Sachs, constitute the majority of NYSE trading volume - and they're trading for their own book, not for customers, will no longer be disclosed."
On the contrary - this is not even close to correct. I invite Denninger to contact me any time he needs an explanation of program trading related items - I'll be happy to explain it to him.

As I explained previously, Broker-Dealers were previously required to file a nightly report with the NYSE called the DPTR - the Daily Program Trading Report - detailing their program trading activities. A "Program Trade" was classified as a trade consisting of 15 or more different stocks, totalling at least $1MM in notional. There are different kinds of program trades: agency (for customers), principal (for the firm) and index arbitrage (capitalizing on differences between the price of futures and the baskets of stocks they represent). In addition, there is "customer facilitation," which is when the brokerage firm commits its risk capital in executing the trade for the client - it takes the risk off the client's books and onto its own books.

Here's what the new rule change from the NYSE does: it automates the DPTR. It's that simple. When program trades are entered into the NYSE systems, they are already coded as "agency," "principal" or "index arbitrage." The exchange is just updating their reporting to take advantage of technology and use this data that they already have, instead of relying on the firms to submit it themselves. It will result in MORE transparency, not less. Why rely on the brokerage firms to submit the data to the NYSE when the NYSE already has the data? That's why they are making the change.

Of course, the headline

"NYSE Uses Simple Technology to Improve Data Quality and Transparency"

isn't quite as catchy as Denninger's "Conspiracy to Hide Bubble Formation."

Brokerage firms like to report large volumes so that they can show their customers that they have the most liquidity and, presumably, expertise in the area. The new rule change will likely reduce reported volumes, as it will prevent brokers from reporting questionable trades such as ETF creations and redemptions on the DPTR. Also, I'm not sure what will happen to the "risk trades" which were previously classified as "customer facilitation." Such trades were somewhat double counted - the trade would be counted when the client traded it to the broker, and again when the broker traded out of the positions.

So why does it bother me so much when someone errs in reporting a simple story? Well, I read the comments on Denninger's site to see how his readers would react. Sadly, they follow like hypnotized sheep. It irks me to no end to read "Well, this is the straw that broke the camel's back for me. I've just stopped contributions to my 401k." Really? You read an article about finance on a website (written by a guy who has expertise in network and computer systems) that's not even close to correct and you're stopping your 401k contributions? Ouch.

Our role as financial bloggers should be to educate people into making smarter decisions - not to scare them into making dumber ones.


note: I have never worked for GS or the NYSE, and have no particular love for either of them. Also I am still short the broad market (SPX), long gold, long silver, and long TBT.


Anonymous said...

Have been long calls and short puts on TBT since January. It's like having an ATM in my living room.

Roffe said...

Thanks for your clarifications!

Anonymous said...

Kid, i have an off topic question and you seem to really know your stuff you im sure you can give me an simple explanation in no time. Question: 3 component index, A B C .33% of index each. A=30vol, B= 50vol, C=80vol. You can make up your own correlations for this example. Could you please show me the steps to work out the index vol for this example using the above numbers?? Thanks a ton.

Kid Dynamite said...

Anon - i'm not really sure how to do the index vol calculation... something with a covariance matrix methinks...

Kid Dynamite said...

anon - stay tuned - i have asked my people and will post in these comments when i get a reply. one guy says he can help...

Anonymous said...

Thanks alot Kid. I really appreciate it. Thanks for taking the time. Please try and do it with the actual numbers. Ive always learned much quicker with actual examples for some reason.

Kid Dynamite said...

here you go Anon - from a vol-slinging friend of mine who needs to remain anonymous:

U first need to calculate the variance of your index and then take the square root to get the volatility.

The easiest way to calculate variance for a three stock index is to build a grid (or matrix) with the three names labeled on each axis. There will be nine empty cells in the grid to start. Have the weights, vols and correlations handy.

For each given calculation, we'll let the stock on the X axis be
"Stock X" and the stock on the Y axis will be "StockY"

To compute the value for each of the nine empty cells you take the
product of weight(stock X) * volatility(stock X) * weight(stock Y) * volatility(stock Y) * Correlation of (Stock X to Stock Y).

You do this for all nine cells, so THREE of those nine cells actually
use the SAME two stocks (so for those you are basically squaring
weight and vol, and correlation is 1.00).

Once you have values in all nine boxes you just sum up the nine
numbers... this is your index's VARIANCE.

Take the square root, and've got your index vol!

Anonymous said...

beautiful, thanks alot kid. really appreciate it!

Anonymous said...

To Bones: I know this is a tragic day for you. The poodle isn't going to be running for el presidente in 2012. I'm sorry. Really, I am. But thats ok. I'm buoyed by the hope that Mitt Romney, who knows less than nothing about anything, will be nominee.....

---Fingers Crossed

Woodshedder said...

Kid, Denninger has written many things before that are just plain incorrect. I took him to task on a statement he made, and he ended up deleting a 4 page thread on his own forum because I proved him wrong.

How's that for honesty and transparency?

Really, he's a fascist, in almost every sense of the word. Ironic, no?

Kid Dynamite said...

Woodshedder: the scariest thing is how his readers take it as scripture, when it's clearly wrong.

Anon - since you didn't use the word "'pubs" i won't delete your post. you have to admit - Palin was ahead of her time with the Bridge to Nowhere stimulus... Pretty soon we'll be building them all over the country.