Wednesday, 17 May 2017

NYSE American - attack of the clones

Today, the SEC approved the application for NYSE's speed-bumped IEX clone.

A hall of mirrors it is:
SEC NYSE Amerian software speed-bump approval
The financial press quickly reported:

This was despite a last ditch pitch for disapproval by IEX,
"On Wednesday, May 10, 2017, David Shillman, Richard Holley, Sonia Trocchio, and Michael Ogershok, all from the Division of Trading and Markets’ Office of Market Supervision, met with John Ramsay, representative of IEX. The discussion concerned NYSE MKT LLC’s proposed rule change to amend NYSE MKT Rules 7.29E and 1.1E to provide for an intentional delay to specified order processing, including the comments reflected in IEX’s public comment letters submitted to date on the proposed rule change."
The SEC had no real choice, given the documents were in proper order, due to the precedent IEX set. This is the sad conclusion BATS' SEC letter also came to,
"However, in light of the Commission’s approval of IEX’s delay mechanism and the Commission’s related interpretation of Rule 611 of Regulation NMS, Bats sees no legal grounds for the Commission to disapprove NYSE MKT’s proposed rule change."
There is a duo of devilish differences to delight us in their deplorability. Let's meander on.

SIP games


Firstly, NYSE American customers will have their quotes and trades exposed on the SIP data feeds before NYSE American reports back to them. That is a little deplorable you'd have to say. This is a similar architecture to IEX but the details are quite different and those differences matter. Here is the relevant piece from the SEC approval:

I have previously talked about the SIP games applicable to the IEX's Dark Fader, "IEX trading." That is, at Mahwah and Carteret you may receive IEX market data before IEX's own local customers. This is due to the CTA and UTP SIP processors being faster than IEX Dark Fader's 350 microsecond delay. Bats' letter to the SEC also pointed out this,
"At a high level, Bats reiterates its position that speed bumps of the nature that IEX employs and NYSE MKT is proposing provide zero benefits to displayed orders 1"
by way of a reference to the same in their footnote,


If we add in the Mahwah|NYSE to Carteret|Nasdaq link and update the latencies to my previous somewhat lame diagram, we get the following approximate sketch:

(click to enlarge)
The SIP processing latencies are medians. I've also added the trade and quote latencies to draw a distinction as the CTA SIP is inexplicably much slower for trade processing. The latencies were drawn from the current reports, except the CTA SIP trade latency which was drawn from the previous report as it showed a more relevant median latency due to the monthly breakdown. Please note Bats Exchange is to be found in the NY4/NY5 complex.

As you may now understand, the latency picture is quite messy when you include the 350 microsecond speed-bumped exchange feeds and order feedback for both NYSE American and IEX into the pictures at Mahwah and NY5 respectively. This gets a little worse when you consider the non-displayed versus displayed aspects of your trading trials and tribulations.

Who benefits? Those who understand the market structure minutiae and also have the resources to expend to put facilities in all the necessary locations. A smart HFT will not like the mess as it acts as a long term friction hampering efficient market development. A smart HFT is also the most likely to benefit due to their laser focus on the small details that ensure their survival. There is a reason why Citadel is often the biggest trader on IEX. That is, an HFT relies on good markets and negative developments, such as these speed-bumped dark faders, are not in the best interest of the market, and, by implication, not in the interest of an HFT.

Supportive HFTs are acting a little sycophantically in my book. They are perhaps disregarding the long-term good for the short-term politic.

You can see by way of the above diagram, for both NYSE American and IEX, simple co-lo is not enough. You need to have processing capabilities in all the main centres if you wish to trade optimally. At least for the CTA Plan's SIP stocks, you should already be colocated for the SIP's 80 microsecond delayed quote feed, around 270 microseconds before other NYSE American customers get their data. It's just more expense.

Do you also find it weird that UTP stocks' data from NYSE American will turn up in Carteret's Nasdaq data centre before it makes an appearance in NYSE's Mahwah facility?

Welcome to an SEC mandated hell.

Dark Fading


The largest part of IEX's success to date in achieving a market share of slightly over two percent is due to its dark trading. Less than twenty percent of IEX's trading is attributable to lit volume.


IEX is a parasitic vehicle that subverts price discovery. Think about that. It is a public exchange that thwarts price discovery, openness, competition, efficiency, progress, and innovation. So far, as IEX has bumped its size up a little, you may see the threatening correlation that is emerging. More market share equates to the likelihood of darker trading:
There is certainly a place for dark parasitic trading. I would argue that place is not within an advantaged public marketplace. The SEC either decided otherwise, or should regret its approval.

Another aspect of IEX's Dark Fader is that it is expensive - very expensive at 9 mills a side. NYSE American may use the same parasitic force to overcome IEX's Dark Fader by simple economics. If NYSE prices its DPEG and Primary Peg equivalents at a more reasonable rate then perhaps the IEX Dark Fader infection may be extinguished by competitive forces. NYSE may price IEX out of existence. Please do so.

This brings NYSE American's own dark fading into sharp relief. Presently NYSE has simply proposed copying an older version of IEX's crumbling quote indicator to power its own dark fading pegs. This will not work as well as it could.

Here, for completeness, are the current IEX Dark Fading formulae, now schmarketed as IEX Signal:

Crumbling quote `if QIF>{(0.39, if spread <= $0.01), (0.45, if $0.01 < spread <= $0.02), (0.51, if $0.02 < spread <= $0.03), (0.39, if $0.03 < spread) :}`

The variable definitions below are quoted from pages 33 & 34 of Exhibit 5 to the March 10 IEX SEC filing. Note that in this filing instead of including all the markets in the number of protected quotations IEX has chosen to incorporate only eight exchanges (XNYS, ARCX, XNGS, XBOS, BATS, BATY, EDGX, EDGA), thus N and F may range from 1 to 8. Three exchanges (XNGS, EDGX, BATS) still get a special mention, as per the last formulae's iteration, in the Delta definition.

  1. N = the number of Protected Quotations on the near side of the market, i.e. Protected NBB for buy orders and Protected NBO for sell orders.
  2. F = the number of Protected Quotations on the far side of the market, i.e. Protected NBO for buy orders and Protected NBB for sell orders.
  3. NC = the number of Protected Quotations on the near side of the market minus the maximum number of Protected Quotations on the near side at any point since one (1) millisecond ago or the most recent PBBO change, whichever happened more recently
  4. FC = the number of Protected Quotations on the far side of the market minus the minimum number of Protected Quotations on the far side at any point since one (1) millisecond ago or the most recent PBBO change, whichever happened more recently
  5. EPos = a Boolean indicator that equals 1 if the most recent quotation update was a quotation of a protected market joining the near side of the market at the same price
  6. ENeg = a Boolean indicator that equals 1 if the most recent quotation update was a quotation of a protected market moving away from the near side of market that was previously at the same price.
  7. EPosPrev = a Boolean indicator that equals 1 if the second most recent quotation update was a quotation of a protected market joining the near side of the market at the same price AND the second most recent quotation update occurred since one (1) millisecond ago or the most recent PBBO change, whichever happened more recently.
  8. ENegPrev = a Boolean indicator that equals 1 if the second most recent quotation update was a quotation of a protected market moving away from the near side of market that was previously at the same price AND the second most recent quotation update occurred since one (1) millisecond ago or the most recent PBBO change, whichever happened more recently.
  9. Delta = the number of these three (3) venues that moved away from the near side of the market on the same side of the market and were at the same price at any point since one (1) millisecond ago or the most recent PBBO change, whichever happened more recently: XNGS, EDGX, BATS.
The parameterisation of the crumbling quote will need to be specialised for NYSE's location and related latencies. It is a fairly straightforward task for the quantitatively inclined, but it is a job that nevertheless still needs to be done. I am hopeful NYSE will take this a step further and produce something a little better and more advanced. The IEX formulae are pretty lame which not only prevents brokers, asset managers, and traders innovating but IEX's lameness acts as a retrograde disservice to the financial community.

NYSE if you need help, my email address is on my contact page.

It could be worse. It soon may be. Let's hope CHX's harmful speed-bump and Nasdaq's rather silly ELO don't add to the NMS hall of mirrors. You'd hope the SEC may see the error of its ways and one day mandate the removal of both IEX's and NYSE American's speed-bumps. Now, that would be a truly beneficial NMS development. The odds are too long to take such a bet. You'd better not wait and keeping rolling out your IEX and NYSE American infrastructure to Mahwah, Carteret, and Secaucus, along with the required microwave, laser, or millimetre wave assets. Life ain't meant to be easy.

Finally, it remains to be seen what attention, if any, NYSE will pay to IEX's patents and patent applications.

Happy trading,

--Matt.

No comments:

Post a Comment