Pity, it is a bad book. Im still very much against HFT.
1) most of the arguments for call auctions instead of continuous trading assume a centralized call auction that in and of itself adds no cost above our current decentralized exchanges. This seems like an extremely troublesome assumption given that we have seen the negative impact of centralized monopoly exchanges.
2) If it isn't a centralized auction there needs to be some price sync. mechanism. The most obvious one seems to be either latency arbitraging HFT (which for some reason people don't like) or a governmental central auction block the type of which is already causing legislative arbitration in the current markets.
3) Call auctions in and of them selves do not solve the major issue with our current prioritization system. If the auction algorithm has several prices that match the same amount of participants, or if the price it determines has an uneven number of buyers & sellers, who gets priority?
4) Price discovery is currently harder in call auction instruments. That could be because they are only done in illiquid markets and liquidity itself will fix that, but it is an issue with the current examples of this style of trading.
5) "Market" order mechanism. Many, many market participants are more worried about making sure they buy/sell than about high fidelity price fairness. If I say, "I want out of Goog no matter what", that complicates any auction methodology and introduces opportunities for gaming.
At the end of the day, I can't figure out who will actually be protected in this style of market and to whose disadvantage. That is a lot of added complexity for a very vague benefit.
And since this is a discussion about a book, it's worth pointing out that the book is focused around an alternative proposal for market design.
They are still a price-time priority matching exchange, they still allow colocation (well not directly, but they are themselves in data centers that have colocation), they still offer exotic order types, and their much hyped added latency amounts to 350 micro seconds. They aren't trying to remove HFT from the markets, they are trying to disadvantage one class of HFT in favor of another.
I personally don't have any problem with their goal, but they aren't being honest about it and that seems problematic.
The stock market is a service to provide equity for real production companies to grow. HFT distorts the real value of a stock. There is also 0 chance of any regulation authority to ever check the amount of paperwork HFT produces.
63.36%: "The NYSE had ... built this 4000,000-square-foot fortress in the middle of nowhere ... the moment they announced their plans HFT firms began to buy up land surrounding the fort ... the NYSE somehow persuaded the SEC to let them make a rule for themselves: Any banks or brokers or HFT firms that did not buy space inside the fort would be allowed to connect to the NYSE in or of two places: Newark, New Jersey, or Manhattan. The time required to move a signal ... undermined HFT strategies ... "There was a precedent: They'd let NYSE do it,"..."
So to my unsophisticated eye, this sophisticated reviewer is blowing a bit of smoke. Not all smoke, but a bit of it at least.
75.73%: "A few thought it was important to remember that technology had lowered their trading costs from what they had been decades earlier - and half-turned a half-blind eye to the stunts Wall Street intermediaries had pulled to prevent technology from lowering those costs even further."
I'm beginning to suspect the author is flat out dishonest. The Goldman cock-up is used as a <i>defense</i> of Goldman. Lewis says that unlike other firms they had a legitimate excuse for not trading with IEX - they didn't trust their systems. So, no the cockup had nothing in particular to do with HFT, but Lewis never implied that it did, and it is impossible to read that paragraph and come to another conclusion honestly. The review author is not shooting straight.