Use access key #2 to skip to page content.

TMFPostOfTheDay (< 20)

Married to High Frequency Trading

Recs

5

October 12, 2012 – Comments (6)

Board: Macro Economics

Author: yodaorange

There was a recent METAR thread entitled: Mysterious Algorithm Was 4% of Trading Activity. [1] LorenCobb commented : For many reasons, we need to reign in the high-frequency traders soon... or all Hell will break loose.

In reading the posts and comments, one thought struck me. METARites and other US investors do NOT realize that they are married to High Frequency Trading. Investors took the marriage vows: for better for worse, for richer for poorer, in sickness and in health. . . till death do us part.

Investors might not recall the wedding ceremony, but it occurred many years ago. It was performed by a legal authority and properly recorded with the county clerk. This marriage is a little different from most US marriages, it can NOT result in a divorce. Annulment is also out of the question. As a practical matter, investors have no choice but to make the best of the marriage. A few points:

1) NYSE specialists are EXTINCT and Jurassic Park is closed. In the “old days” all orders for NYSE stocks ended up at a specialist post on the NYSE trading floor. The specialist was tasked with maintaining an “orderly market” for each stock. The specialist matched up buyers and sellers. The specialist was also the buyer and seller of last resort. So if no buyers had open bids, the specialist would step in and buy the stock for “his book.”

Buyers and sellers paid a price for this service in addition to brokerage commissions. The price was the “bid-ask spread” and was typically in the range of 1/8th (12.5 cents) up to ½ (50 cents). The spread depended on several items, but in the end, the customer would end up paying TWO spreads in order to buy and sell an issue. The specialist business was lucrative. Specialists would tell you that they lost money about 2 to 3 days per year. They made money on the other ~ 237 trading days. Not a bad business model.

There were 35 different specialist firms on the NYSE in 2000. Today there are 3.1. The 0.1 firm is Knight Capital that recently blew themselves up with an errant trading program. Specialists are now called “designated market makers” aka DMM. The largest DMM is a Chicago based company named GETCO. GETCO is also one of the largest if not the largest High Frequency Traders. The other two specialist firms are Goldman Sachs and Barclays.

In the case of GETCO, they now have the official NYSE responsibility of maintaining an orderly market while at the same time trading for their own accounts with their HFT.

IF HFT WAS SOMEHOW BANISHED TODAY, THERE ARE LITERALLY NO HUMAN SPECIALISTS READY, WILLING AND ABLE TO RESUME THEIR OLD ROLE.

Small investors have benefited from the transition from specialists to DMM. Spreads have shrunk dramatically over time. It is common to find actively traded issues with 1 and 2 cent spreads compared to the 1/8th to 1/2ths that used to exist. Granted, the spread is NOT an issue to the few long term buy and hold investors left.

2) Exchanges EXIST for HFT’s not small retail investors. The NYSE trading floor at Broad and Wall in New York City has become a stage prop. Its main value is for photo ops. The actual NYSE trading floor is in Mahwah, New Jersey. It is a large building, with Fort Knox level security that is home to the NYSE computers PLUS the “co-located” computers used by the DMM’ and HFT’s.

Do you think the NYSE said: “We need to build a new state of the art facility with ~ zero humans present to improve trading for the small investor?” NOT A CHANCE.

The exchanges have added many different “order types “strictly for the benefit of HFT’s.[2] The small investor does NOT use any of these. My broad opinion is that these order types are for the benefit of HFT to the detriment of the small investor. Rimpy will correctly point out that I have NO proof that these are to the detriment of the small investor. Rimpy also has NO proof that these order types are NOT specifically to get an advantage for HFT’s.


IF HFT WAS SOMEHOW BANISHED TODAY, THE NYSE AND OTHER EXCHANGES WOULD SEE AN IMMEDIATE DROP IN SALES VOLUMES AND PROFITS. THEY ARE NOT GOING TO WILLINGLY CUT THEIR VOLUMES AND PROFITS.

3) Congress stands ZERO chance of making meaningful changes to the law that would reign in HFT. On September 20, 2012 the Senate Committee on Banking, Housing and Urban Affairs held a hearing Computerized Trading: What Should the Rules of the Road Be? [3] They called four industry experts to testify. You can download and read the prepared testimony of all four. I read them all.

I did NOT post anything on METAR at the time, because my main point was simple:

HFT’s are from Venus and Senators are from Mars. There is no way on God’s green earth that these Senators could understand HFT’s and/or implement meaningful reform. I don’t care how smart the Senators are in their own field, but NOBODY can take a 1 hour short course on HFT’s and make meaningful and CORRECT reforms.

Matters of fact, IF Congress were to tackle HFT reform, chances are good, they would make the situation worse instead of better. I WOULD NOT COUNT ON CONGRESS SAVING THE DAY ON HFT REFORM.

I could write a short story on this Congressional hearing but don’t think it is worthwhile

4) SEC. . . is CLUELESS regarding HFT. I posted recently about the importance of “knowing what you don’t know.” Unfortunately, Mary Schapiro and the SEC staff must have missed that post. Many of us have been in situations where we said to ourselves: “I am the dumbest person in the room and I need help understanding this topic.” When you find yourself in that situation, there are several possible responses:

a) Ostrich- Stick your head in the sand and hope the problem solves itself.

b) Volunteers- Ask your internal team who would like to volunteer to become the expert on this topic and lead the effort.

c) White flag- Raise the white flag and hire the absolute, best, most qualified people OUTSIDE your organization to get you up to speed.

The SEC’s original response was the Ostrich plan. It worked OK from 2000 through 2010 until the Flash Crash in May. Then the SEC went out and hired on a contract basis Gregg Berman who has worked in the HFT industry and has a physics PHD from Princeton. Clearly a very bright fellow. The SEC published a report on the Flash Crash. [5] I posted a critique of it entitled SEC Lays an Egg on Flash Crash. [6] Short summary is that I thought the SEC was way off base in their analysis.

The SEC also had the brilliant idea of building a multibillion $ computer system from scratch to log every piece of market information from every exchange, dark pool, etc. Luckily they dropped that idea and signed a contract with an existing HFT company, Tradworx, to build the system for them.[7] They announced the contract on July 20, 2012, a few billion $’s short and a decade too late.

On October 1st, 2012 the SEC hired Gregg Berman as a full-time employee to START a new Office of Analytics and Research. This is progress, once again it is 2 or 5 or 10 years TOO LATE.

The problem with all of this is that the SEC has TOTALLY IGNORED the outside team that could help them TODAY. That firm is the “garage shop” Nanex. It literally is a few guys, a small office and some serious computing hardware and software.

The Nanex analysis of the flash crash was INFINITELY better IMO than the SEC’s analysis. One of the root causes that Nanex identified was two exchanges that got out of time sequence. One exchange thought it was time zero. Another exchange thought it was time zero plus 1 second. You can NOT mix prices from different times otherwise all Hades breaks loose like occurred during the flash crash. (Obviously I have simplified this explanation. See Nanex.net for infinitely more detail.)

At a minimum, the SEC should have contacted Nanex to hear their view of what occurred. Did the SEC contact them: NO. Has the SEC contacted them subsequently: NO. Matter of fact, the internal story is that the SEC strongly dislikes Nanex because it made the SEC look like a bunch of fools. The SEC should have admitted that Nanex was light years ahead and sought their help. Unfortunately, the SEC was too prideful to admit how far behind they were and decided to re-invent the wheel. BIG MISTAKE.

Bottom line on the SEC is that I would not look for them to make any meaningful reforms to HFT. The flash crash occurred on May 6, 2010. Can you name any and all of the reforms the SEC has implemented since then? The main changes were a new rule that allowed trades greater than 30% from the market to be “busted.” 29.99% trades are OK. The other changes are adding “speed bumps” which are somewhat after the fact. ZERO has been done to proactively PREVENT another flash crash type event.

BTW, did I mention that the large HFT firm GETCO has hired several high level SEC staffers in the last few years. They hired Elizabeth King on June 11, 2010 about one month after the flash crash. [8] I am sure the timing was coincidental. King was associate director of the SEC's trading and markets division, which is responsible for supervising exchanges, brokers and credit rating agencies.

5. “Perfect” changes to HFT are NOT obvious. It would take another short story, but in simple terms, there is near unanimous agreement that all of the proposed HFT solutions will cause other problems. Stated differently, we aren’t in Kansas anymore and we don’t have the red slippers to click three times on. There is also increasing agreement INSIDE the HFT industry that the situation is precarious and the system is marginally stable.

6. How and when will HFT change? Most likely Congress will get involved AFTER we have another flash crash type meltdown or worse. Let’s assume they have a timely response like they did to the 2008 Credit Crisis. Here it is 2012 and their response was Dodd-Frank, which did ZERO to permanently solve the TBTF banks. Need I say more? Probably a few years AFTER the next major event, some new law will be enacted that addresses the wrong problem with a wrong solution.

BOTTOM LINE: is that all US investors are married to HFT for better or for worse. Better get used to the marriage, because NOTHING is going to change any time soon.

Thanks,

Yodaorange

[1] Desertdaveatol post: Mysterious Algorithm Was 4% of Trading Activity
http://boards.fool.com/mysterious-algorithm-was-4-of-trading...

[2] Wall Street Journal article: “For Superfast Stock Traders, a Way to Jump Ahead in Line”
http://online.wsj.com/article/SB1000087239639044398920457759...

[3] Senate Committee hearing on HFT
http://banking.senate.gov/public/index.cfm?FuseAction=Hearin...

[4] SEC contracts with Gregg Berman to issue report on the May 2010 Flash Crash
http://us.mg4.mail.yahoo.com/neo/launch?.rand=4tl90b3rioq9s#...

[5] SEC report on Flash Crash
http://sec.gov/news/studies/2010/marketevents-report.pdf

[6] Yodaorange post: SEC lays an Egg on Flash Crash
http://boards.fool.com/sec-lays-an-egg-on-flash-crash-287997...

[7] SEC contracts with HFT firm Tradeworx
http://www.tradersmagazine.com/issues/25_342/sec-trading-dat...

[7] SEC Hires Gregg Berman to start Office of Analytics and Research.
http://www.bloomberg.com/news/2012-10-01/sec-leads-from-behi...

[8] GETCO hires SEC administrator Elizabeth King
http://www.reuters.com/article/2010/06/11/us-sec-getco-idUST...

6 Comments – Post Your Own

#1) On October 12, 2012 at 11:32 AM, Mega (99.95) wrote:

You can do market making without false quote stuffing.

Not really sure what your point was, this post was way too long.

Report this comment
#2) On October 14, 2012 at 12:49 PM, awallejr (82.72) wrote:

I think that a bit harsh and off point about the post's length.  I thought it well presented and argued.  I confess that my eyes tend to cross once the actual "science" of HFT is discussed, but it does concern me.

While the SEC has the jurisdiction, they just don't have qualified people.  It took a major market crash to uncover Madoff.

I am curious tho if re-instituting the "uptick" rule would have any worthwhile value.  Would it kind of put the computers on hold for a second?

Report this comment
#3) On October 14, 2012 at 3:38 PM, ADrumlinDaisy (< 20) wrote:

"this post was way too long"

 Yeah --  and it used way too many big words too.

 But I did get a smarter friend to read it for me, and he told me it was a really interesting post!

Thanks!

Rich

Report this comment
#4) On October 14, 2012 at 7:24 PM, NajdorfSicilian (99.88) wrote:

HFT has been around for 25 years. It's the order stuffing that needs to be fixed, that's all.

Report this comment
#5) On October 15, 2012 at 11:09 AM, captainccs (< 20) wrote:

As an individual investor I don't think HTF has any effect on my trades

Report this comment
#6) On October 17, 2012 at 4:40 PM, kbtoys99 (< 20) wrote:

HFT is a real problem. It will ultimately lead to another flas crash if we are not careful. there are no "real" bids or asks out in the market anymore. just algo daytraders.

Report this comment

Featured Broker Partners


Advertisement