Skip to main content

Michael Lewis: Flash Boys

Michael Lewis is a contemporary of mine (no joke didn't know him but he did his MA in Economics at the LSE at the same time I did).  Granted he did much better than I, and is a man of conviction.  ML started in Salomon Brother's bond sales desk in the early 1980s -- Although the world believes that Salomon was a investment bank (it was in some small scale) where it dominated was in the bond market.  In fact, long after ML left the bank it was convicted of cornering the US gov't short term bond market -- it takes some massively large brass balls to try that (even in the 1980s).  Salomon Brothers was the biggest bond house on Wall Street.

Anyway, ML wrote his first book:  Liar's Poker that was a cautionary tail as to what banks actually did -- the amazing thing (and it shows the level of depravity of young bankers) was that future bankers used the book as a "how to get into banking manual".  ML wrote a number of other great books (Blind Side, Money Ball, The New new thing) and some less good books (the Big Short: story was great book was too long).

Anyway, last night Amazon started selling ML's new book:  Flash Boys that talks about the High Frequency Traders (HFT for short), something that started about a decade ago (early for some but not in the institutionalize way you have today).  What the HFT guys need is fast computers (algorithm ) high speed bandwidth and cheap execution cost.

I've not read the book yet, but HFT do two things; first they trade ahead of the market -- as is explained in the 60 minute clip they are able to execute a trade ahead of the rest of Wall Street.  The second thing (which I guess he talks about in the book) is price discovery.  Price discovery is very important it tells you where the demand/supply curve really is.  HFTs do that by flooding the exchange with orders -- and removing these order almost instantaneously -- but slow enough to trigger actions by the "regular" market participants (which generates a "failed trade"), this way the HFT discover at what price institutional money start to sell stock and at what price options will trigger  activity (buying/selling/exercising).  Having that information allows you to "game" the market.  Knowing how the price will affect the market behaviour is a huge structural advantage -- in fact the HFT don't care about the price they just know the present a few seconds ahead of everyone else, and they know what the market will do when the price changes!

You know the problem is serious when Fidelity, Goldman and a bunch of massive hedge funds (multi billion) are concerned about the problem -- the attractiveness of ML's book is that it puts it all in perspective. BTW the real question is why do exchange allow HFT traders to operate -- the reason is revenues: HFT account for about 50% of all exchange revenues -- a source of revenues that did not exist a few years ago.

Finally, how much are we talking about?  Well imagine one HFT spent $300 million in optic cable to shave 3/5 milliseconds from its trading.  That's a lot of money for such a small time advantage.

The other attractive aspect of the of the book is that it shows the solution to the problem... which was the creation of a exchange that stops the HFT from functioning, the method is very interesting, but HFT are really really smart and have huge incentives to find a solution.  One thing for sure is that the HFT's form of "legal" frontrunning may become more difficult execute.  Don't know what they are going to do about price discovery.

Anyway, interesting stuff

Note:  BTW this is not saving the market for the little guy -- there is no little guy, the issue here is that the markets are rigged by some of the institutional participants (like the LIBOR scandal of last year).  This is not about Every Man Trader; its the Big Boys who are complaining; hence the Shadenfreude


Popular posts from this blog

Trucker shortage? No a plan to allow driverless rigs

There are still articles on how America is running out of truckers -- and that its a huge problem, except its not a problem, if it was a problem salaries would rise to so that demand would clear. Trucking is one of those industry where the vast majority of participants are owner/operators and therefore free agents.

Salaries and cost are extremely well know, "industry" complains that there are not enough truckers, yet wages continue to fall... Therefore there are still too many truckers around, for if there was a shortage of supply prices would rise, and they don't.

What there is though is something different; there is a push to allow automatic rigs to "operate across the US", so to encourage the various authorities to allow self driving rigs you talk shortage and hope that politicians decided that "Well if people don't want to work, lets get robots to do the work" or words to that effect.

This has nothing to do with shortage of drivers, but every…

Every punter says oil prices are on the rise: Oil hits $48/bbl -- lowest since September 2016

What the hell?

How could this be, punters, advisors, investment bankers all agreed commodity prices  in general and oil prices in particular are on the rise...its a brave new era for producers and exporters -- finally the world is back and demand is going through the roof, except not so much!

What happened?  Well energy is complicated, the world operates in a balance -- 30 days of physical reserves is about all we've got (seriously) this is a just in time business.  So the long term trend always gets hit by short term variations.

Global production over the past 12 months has risen by somewhat less than 1.5% per annum.  As the world market changes production becomes less energy intensive (maybe), but the reality is that the world is growing more slowly -- America Q4 GDP growth was around 1.9% (annualized) Europe is going nowhere fast (the GDP growth in Germany is overshadowed by the lack of growth in France, Italy, Spain (lets say 27 Euro members generated a total GDP growth of 1.2…

Paying for research

This morning I was reading that CLSA -- since 2013 proudly owned by CITIC -- was shutting down its American equity research department -- 90 people will be affected!

Now the value of a lot of research is limited, that is not to say that all research is bad. In fact, I remember that GS's Asia Aerospace research was considered the bible for the sector.  Granted, there was little you could do with the research since the "buy" was for Chinese airlines...that were state owned.  Still it was a vey valuable tool in understanding the local dynamics.  It seems that the US has introduced new legislation that forces brokers to "sell" their research services!  Figures of $10,000 an hour have been mentioned...

Now, research can be sold many times; if GS has 5000/6000 clients they may sell the same research 300x or 400x (I exaggerate) but this is the key -- Those who buy the research are, I presume, prohibited from giving it away or selling it, at the same time the same rese…