Wednesday, 29 January 2020

Out of touch

I spent most of December in critical care with a family member. Medicine is more voodoo than markets. Death, misdiagnosis, errors in surgery (aka cut nerves), months of incorrect treatment. It's a silly world. The last eighteen months could have been better. It continues for now.

Today was a day procedure, so I drove up the mountain behind Hobart whilst the patient was under a general. A bit hazy but a pleasant place to look over the burning, infected planet.

Fuzzy Hobart today, Jan 28th 2020 [A2 phone]

I saw Mr Kipp Rogers chimed in on the "HFT Arms Race" FCA paper by Matteo Aquilina, Eric Budish and Peter O’Neill with his note, 'A Quick Note About the FCA Occasional Paper on “Latency Arbitrage”'. It's a kind note from Mr Rogers. The Aquilina paper has some interesting features. I do find it joyful that the academic world continues to fumble the ball when considering markets. At least there should be a place for me if I can get back to trading this year even without suitable a PhD. It doesn't matter how many footnotes or referees, no one seems to get it quite right. This is true both of the papers I disagree and I agree with. We all tend not to call out the papers we like the sound of though.

Five hundred microseconds is an eternity in such a world and you need to think about what you really are seeing if you're seeing features at that scale. Consider taking the expensive direct market data feeds in Chicago and NJ, say for ES and the ye olde mutant eight-legged ETF. Ask yourself how does the synchronisation between those two markets seem to happen faster than the speed of light? The multi-faceted answer to that question makes you realise that such papers are deeply flawed on more than one axis, particularly in their labelling of features with names bearing little relationship to the implications of the measure. Such mislabeling, or bias, projects heat rather than light into what may be an otherwise fruitful discussion regarding interesting indicators or measurements.

You could say the same for yesterday's WSJ report by Mr Alexander Osipovich, "Ultrafast Trading Costs Stock Investors Nearly $5 Billion a Year, Study Says." Mr Osipovisch reports the paper fairly. However, again, this underlying paper is a misdirection that attributes supernatural powers to market data that is a little improper. The measurements should lead to useful discussion but to call it as "costs stock investors nearly $5 Billion a Year" is unhelpful. More light and less heat would be nice.

Bashing HFT has long been a worthwhile click unfortunately since the gross fiction of "Flash Boys." Articles get clicks. People like me write notes on it. The cycle continues. Not very productive.

Asia Pacific HFTs' business is easing a little but it continues to be good. Large private companies in Australia are publicly reported on so we get to see how the bigger HFT firms are doing here, though lagging by some 18 months or so. The most recent, though old, figures for the 2017/18 year of "total income" were released in December:

Susquehanna Ireland Limited   $166M
Susquehanna Pacific Pty Ltd $324M
IMC Pacific Holding Pty Limited $265M
Optiver Australia Holdings Pty Limited $424M

The previous reporting I noted here FWIW.

People object to profits. This drives some of such anti-HFT news cycle sentiment. HFTs are winning so they must be evil. HFT's can't have systematic losing days by definition as then they'd be out of business. HFTs continue to bear the bare cross of making less money than the market participants they creatively destroy. In the same ATO report, Deutsche scored $1,380M of "total income", CS has $820M, and UBS shows $1,390M. You should worry more about those behemoths. Cheer the little plucky HFTs as the HFTs will take income from the large institutions and make less so you can make more. That's how it is meant to work.

As IEX fades I see it is trying to put forward a D-limit order. This is the kind of improper market structure people feared public markets would lurch towards when IEX originally applied for its license. This should be a Sisyphean moment, but I fear it is not. Short memory it seems.



Midnight Oil with an old favourite from 1983

And whilst US markets continue to deny the need for variable and sub-penny ticks you get the virus of PFOF continuing to both benefit retail and deny best-execution. You'd hope the easy and obvious decisions would be made but then you look at American politics and realise you have to play the hand you're dealt.

Apologies for continuing to be MIA. I hope to get back on track in March or April if the voodoo levy breaks.

Happy trading,

--Matt.

No comments:

Post a comment