Quality markets have always had, and will always have market makers. These market makers provide a service to the market and are compensated for the risk that they take. In years past, the specialists of the NYSE served as market makers. Today, HFTs are market makers. To say that HFTs extract money from the ultimate exchange between buyer and seller is true, but demonstrates a lack of understanding as to how markets operate.

What benefits to anyone but themselves? HFTs extract money from the ultimate exchange between a genuine buyer and seller. How is that even conceptually good for either of the muppets?

The chart embedding does not seem to work [ though I was able to see it on the NYT website ].

I am going to make a number of points here:

1. The NYT chart does not support the Nanex chart in any meaningful way: Nanex seemed to imply that quote volumes have gone up 15? 20? times since 2007, while the growth in the NYT chart is much smaller. The NYT chart also measures quote traffic in the single busiest minute; it is a reasonable assumption that the average, or the 90th percentile, would have increased much less than the numbers for the busiest minute.

2. There are two very good reasons for quote volumes to be up since 2007, beyond the (further) proliferation of HFT:

- there were simply more news driven moves in that period than in any other recent period, and news driven market activity happens in the minute or two after news breaks;

- the continuing growth of market centers is leading to greater numbers of both liquidity providing and liquidity taking orders. For example, if all you had was the NYSE, you would send your order for 1000 shares only to NYSE; now, you have to split it into 100-200 share chunks and send to NYSE, ARCA, NASDAQ, BATS, EDGA, EDGX, …. The trading volume does not change, but the quote volume goes up a lot. Many HFTs run strategies designed to exploit this fragmentation, and many couldn’t care less if the fragmentation went away.

I can say with a fair degree of confidence that HFT in the US equity markets pretty much reached its peak in 2008. No significant new participants have sprung up since then, and no existing participants have greatly expanded their activities.

3. You also allege that HFTs use quotes to mask their true activities. This is, to put it bluntly, absurd. Although people call them “quotes”, the stockm arkets are not quote-driven but order-driven systems. And each order is a commitment to trade at that price in that quantity. These orders are churned a lot, but I will submit that most of this churn happens because it can: an HFT system trades on a signal, it recalculates that signal as often and as fast as it can, and it submits or cancels orders based on what the signal says. A lot of this submission or cancellation happens, in all honesty, because there are no constraints on how much of it you can do, or how fast. Once you place reasonable constraints, in the form of minimum trade-to-order ratios, or minimum resting times, the churn will go down. That doesn’t mean HFTs are engaging in any nefarious practices now which they will not engage in then.

4. You echo the NYT in saying that the cost benefits have been through the narrowing of spreads, and that this has stopped around 2008. While I broadly agree with that assertion, I will also say that another big cost reduction has been through a decrease in brokerage fees, which has come about mostly through automation and the adoption of algos/HFT technologies.

PS: if it is not clear, I work in the area. I’m happy to discuss any of this at greater length if you wish.

as the head of electronic trading at a major sell-side shop asked me last week, rhetorically, when I made a similar comment to Tabb’s about fragmentation being the problem:

“how do you put the fragmentation genie back in the bottle?”

the reason markets got fragmented (different places to execute your order) is precisely because we got rid of the NYSE specialist system (which, as the NYT article notes “No one is saying that we go back to the floor specialists”) and replaced it with *competition.*

So it’s important for people to understand: the fragmentation of markets – the competition amongst different market centers to offer the best price and liquidity – is ITSELF what leads to an entire subset of HFT spewing all of these quotes to make sure that they 1) maintain their place in line and 2) maintain pricing order across the fragmented exchanges.

Is there further “benefit” to this for the rest of us? Maybe not – I see your point, but my counterpoint is: there’s also no reason to run around screaming that the sky is falling and that the robots are taking over, or that “the entire market could blow up at any second”.

Who cares – seriously – it really truly doesn’t have any affect on you when the algo-bots quote themselves to death. fact: You can’t even see it!

Aside, I am definitely guilty of generalizing that the alternative to our current structure is the NYSE Specialist monopoly which is terrible for investors and traders.

random thoughts: I think that most people advocating for a transaction tax have no idea what they are talking about, and that the consequences could be dire and are definitely bad for everyone. I also think that, as you mentioned, there is little/no benefit (and also, perhaps, little/no harm!) to continuing to try to speed things up on a scale that retail investors cannot even see (as I already noted), and that a limit on speed/quote volume probably wouldn’t hurt bid/ask spreads too much (ie: no more than 5 quoting/trading time periods per second – just throwing 5 out there as a random number – others have suggested various numbers from 1 to 100)