Rage BY the Machines
BIS on FX HFT – “No Problem”
There was a time that I hired a good number of FX and options traders. I was good at spotting those who had the “right stuff”.
This was not then (nor is it now) a matter of a superior education. Gender was not a big consideration either (I preferred women, as the macho, big balls thing gets in the way). I never found that a person with an economics major had an edge. The fact is that the rational thinking behind economics has little to do with the day-to-day functioning of markets.
I looked for decisiveness. I would ask stupid questions in the middle of an interview:
What’s your favorite color?
Do you eat Jell-O?
Do you like jazz?
If anyone answered theses types of questions with these; they didn’t get a job:
Well, I have several colors that I like. Or;
Err, sometimes I eat Jell-O, but only when I’m sick. Or;
Yes, sometimes I like jazz, but I prefer rock and roll.
The “correct” answers are, Red, No, and Yes. It was irrelevant whether they actually liked Jell-O, but, they had to have an immediate black or white response to a black and white question.To be effective as a trader required the capacity to make near immediate responses to questions. An FX trader who has to think a few minutes to sort out facts and new information was never going to survive. The ability to immediately answer the question: “Sell or Buy?” was a key factor in my hires.Now jump forward 20 years and consider the role of HFT trading in FX. What I thought was a critical variable (speed of decision making) has been brought to a level that I never considered possible. This from a BIS report on FX HFT yesterday:
HFT participants in FX can operate with latency of less than one millisecond, compared with 10–30 milliseconds for most upper-tier non-HFT participants (for comparison, it is said to take around 150 milliseconds for a human being to blink).
The robots that are making FX prices are doing so a rate 150 times faster than the blink of an eye. When hiring “real” people I was looking for fast (2-second) Yes or No answers. Computers are now doing it 2,000Xs faster than that. While that should not really be a surprise given what we know about HFT trading for stocks, this information still blew me away.
The conclusions of the BIS report are similar to the thinking regarding HFT and its role in the stock markets:
HFT helps to distribute liquidity across the decentralised market, improving efficiency, and has narrowed spreads.
The study concluded that of the $4 Trillion of daily FX turnover machines are now executing approximately $1 trillion. That’s a hell of a lot of money.
The study does point out the obvious flaw in HFT participation in the FX market:
HFT market-makers have no binding obligation to stay in the market and place quotes in an adverse market environment.
Two examples were provided to back up this observation. The first, a flash crash of the EURUSD on May 6, 2011 the second, the flash crash of USDYEN on March 5, 2011. The results are pretty clear. When times get tough in the FX market, the robots that were “supposed” to be providing the liquidity went on a coffee break. The additional conclusion was that very high levels of HFT trading preceding the crash contributed to volatility. They used this chart to make the point.
The BIS concludes that HFT is not a big concern for the FX market:
There are key differences in market structure that may make a flash crash-type event less likely in FX than in equities
Less likely? What does that mean? The report points to two flash crashes in FX in 2011 (and it’s only September).
I think the study was built on a false premise. The BIS thinks that the May 6, 2011 flash crash in the US stock market was the result of Waddel & Reed. This has been discussed many times before. This is the conclusion of the SEC. I think it’s totally false. BIS states:
In general, systemic risk is perhaps more likely to be triggered by a “rogue” algorithmic trade than by pure HFT activity.
This is certainly consistent with the emerging evidence about the 6 May flash crash in US equities, which seems to have stemmed from a large sell order executed through an algorithm.
If you start with the assumption that the BIS study is flawed (by virtue of the incorrect conclusion that Waddell&Reed was the cause of the crash) then you have to also assume that the BIS is incorrect on their conclusion that FX HFT is benign.
That’s my conclusion. I think HFT adds to instability. If I’m right, we’re going to see more “Pocket Drops” in FX markets in the months to come. There is no mechanism to DK trades in the FX market (as has been done over at the NYSE). One day some big player is going to get conked on the head in a matter of milliseconds.
HFT (FX or stocks) exists for one reason. It’s profitable business.What benefits it may bring for narrowing spreads and increasing liquidity in normal times is far outweighed by the fact that it’s also the cause of/contributor to instability when the going gets rough.
There is every reason to believe that some “rough going” is ahead of us. I’m looking for big spikes in FX Vol. With that will come some big name casualties.