Monday, 19 February 2018

The market factory: things that make you go hmmm

This little meander is just a catch-up on things that make me go, "Hmmm."

A belated HFT St Valentine's thought first:
Heatsinks are red
Packets race through
Bits don’t worry
Neither should you

CME Canaries

Via the WSJ, Quantlab started a very public war with CME and some other traders, likely to be other HFT firms, who have been enjoying the suffocation of canaries as meandered here, "Killing canaries.

With ten percent of trades having two milliseconds of delay in the market data feed, Quantlab has a good deal of right on its side.

Unfortunately, for avian-friendly traders at CME, Quantlab's representation in the WSJ was a bit wrong. The delay is more than twice as bad as reported. Or twice as good if you like killing canaries.  Quantlab posed to twitter more pertinent statistics,

That is quite the problem. At 5,093,000 nanoseconds of canary, Alexandre Laumonier, aka SniperInMahwah, pointed out - though with terser language - you can get a canary, package it up, transport it from Chicago to your choice of data centres in New York, New Jersey, and Mahwah, execute a trade, all whilst pouring a cup of coffee, and other participants at CME don't yet know the market price has changed.

I don't have a problem with little canaries. The canaries are natural and all exchanges used to work this way. A canary may be thought of as a technological rebate, as compared to a price making rebate, where you get a little advantage for making a market. You may argue for and against. Especially as there is a little more to the story that an HFT may not disclose. You know, secrets. We have to make a living.

Many exchanges are moving to a "public first" position, but I reiterate, having small canary effects is not an evil in itself. A friendly punch in the arm in jest is OK. Giving someone a black-eye is not so great. 5,093,000 nanoseconds of canary is also not OK. IEX remains the worst exchange as they purposefully notify all other co-lo facilities of all trades before their own customers.

I added an addendum to the previous meander noting that the wise Mr Sam Tyfield noted that under EU regs canaries on both lit and dark venues may be termed "phishing" and would be improper unless you documented your lack of intent in a valid trade plan. To not do so would be to invite risk. Mr Tyfield also contributed a nice poem for St Valentines day:
Canaries are contentious
Feelers are too
And if you want to know book depth
No intent is the clue
It seems to me that we all owe thanks to Quantlab for raising this debacular debate. The momentum is certainly swinging toward freeing canaries from market mining cages.

Canary latency is one of those things that should make you go, "Hmmm."


My meandering mind had mixed thoughts around the disbanding of the Equity Market Structure Advisory Committee (EMSAC) in January just as the first Fixed Income Market Structure Advisory Committee (FIMSAC) met.

My first thought was that there is plenty still to do with respect to equity markets. So, why disband? However, as I noted in "U.S. Equity Market Structure Part I: A Review of the Evolution of Today’s Equity Market Structure and How We Got Here," it seemed a futile body with little consensus, competing agendas, and the dialogue often appeared a little irrational. Themis Trading's Mr Joe Saluzzi was an unusual EMSAC inclusion due to Themis Trading's tiny size, his pronouncements being a little weird from time to time, and also as he is prone to shoot from the hip without enough reflection. All qualities that may similarly dismiss consideration of myself from such a post ;-) For example, Mr Saluzzi was reported in the FT saying that the IEX rebate cleverly disguised as a rebate was not a rebate and it was only applicable to principal order flow, in opposition to IEX's clear documentation:

Not an isolated incident.

In the last EMSAC, IEX's Mr Brad Katsuyama made some rather offensive, ridiculous, and wrong statements about price-providing rebates which I discussed here, "Rebate trafficking." As many know, I'm not a fan of IEX's ridiculous posturing. The evil Dark Fader exchange preys upon others' good price discovery work by subverting market efficiency whilst ripping their clients with ridiculously high transaction fees as they send canary-like advance notifications of their clients' trades to Nasdaq and NYSE in advance of their own clients whilst denying their clients the fairness of a colocation environment. Ahhh, that feels better. As a public exchange, IEX sux. You should go, "Hmmmm."

However, the "other side" at EMSAC could also be found foot faulting. I'm generally supportive, admiring even, of Global Trading Systems and the work of Mr Ari Rubenstein. Pushing his agenda as a designated market maker, Mr Rubenstein argued for his vested interest and against the introduction of closing auctions at other exchanges, specifically at CBOE/BATS. That is not very defensible. Such competition is healthy. It is not quite as simple as more competition is better as there are complications around having multiple closes, but it is nevertheless an obviously good move. It goes to the point that EMSAC was a little dysfunctional. The committee format really does nothing that can't be done by canvassing various types of market stakeholders at a leisurely pace. At least the SEC took the sensible path and alternate market closing auctions are now a thing.

EMSAC's dysfunction was also highlighted to me by Instinet's Mr John Comerford's presentation. He had the most obvious and uncontroversial proposal. There is an undeniable economic argument which Mr Comerford succinctly laid out: market tick sizes should be variable. The market microstructure economics are fairly clear and beneficial. There is a general consensus around this issue with Nasdaq certainly supporting this too. Mr Comerford came to the committee with this small and uncontroversial idea quite cleverly trying to bypass contentious issues that had little chance of traction. This variable tick idea is a market aspect that should be a reasonable and straightforward thing for EMSAC to get behind. It seemed, to me, to be simply drowned out by the surrounding noise as little has been done. If EMSAC couldn't push forward something so simple at the SEC, being disbanded is a reasonable outcome. It will be interesting to see if FIMSAC is also broken.

It certainly seems the SEC needs to rely more on its own analysis and make decisions by ignoring the noise where sound economic arguments and clear public interest exists. Seek counsel, advice, comments, but not consensus. That's a thought. Hmmm.

Show me the data

The common refrain, "show me the data" is often used in a market debate to shut down the other side's argument. This is a line that is often valid, but equally makes me go, "Hmmmm."

We have strong economic and financial theories which often need little debate for particular circumstances. If you prove mathematically A + A = 2A you don't need data to reassure you. The standard for when data is or is not required is quite the fuzz in markets. We should all be aware that the "show me the data" mantra is often just harmful.

It reminds me of the common open source refrain to someone pestering you for fixes, "you have the source, send me a patch." Sometimes that is an honest wish, but more often it's just a successful way of shutting down debate. You know you'll never hear from them again.

In data we trust. Except when you see your first black swan with white tipped wings in the river by my house in Huonville. I'm not sure Taleb has seen this variety but he would not be surprised. Inductive hypotheses with misconstrued incremental steps can be dangerous with data too. Data sometimes misleads us. Data often confounds us with concurrent experiments, policies, phases, reactions, assimilation, adjustments, and experiential learning feedback contorting us to consider it may be turtles all the way down. Such data review may be as useful as observing that popular butterfly's wings for distant tornado prediction. It is difficult repairing aircraft whilst they fly. 

I like to look at data, but I don't always need it to know what is right. Too often it's a crutch. Next time when someone challenges you with, "show me the data," don't dismiss the call, but do go, "Hmmmm."

Market data fees

IEX is a public parasite that deserves to be kicked back into ATS land where it belongs. One good thing IEX does, however, is offering a rebate to all of its market participants in the form of free market data. IEX has a lot of rebates for an exchange that offers no rebates. Such a rebate is to be commended. It is not necessarily the most useful market feed as, being a little lame, IEX participates rarely at the NBBO, but this market data rebate is unequivocally good. 

"In 2008, two years after becoming a public company, NYSE began increasing fees for non-core data and the other exchanges quickly followed suit. Despite SIFMA and others expressing concern about these fees, the Commission has allowed the exchanges to continue raising prices on proprietary market data, thus allowing market data to be a primary source of income for the exchanges. Exchange revenues in 2016 were driven by a 29.2% increase in market data revenues to a total of $5.4 billion."
Being such an important part of exchanges' income, market data fees are a controversial issue for exchanges.  BATS offered free market data at its inception. Eventually BATS started charging participants for the service. Mr Dave Cummings, BATS founder and ex-CEO, has publicly expressed disappointment in BATS choosing to charge for market data.

Often participants complain that it shouldn't be right that exchanges are just selling them their own data back to them. I kind of enjoy that argument. It brings a smirk. But, it is not a valid argument. It is a bit like the atom defence in patent litigation that nothing should be patentable as it is just rearranged atoms and we know all the atoms. There is value in the aggregation, processing, and careful delivery of market data.

A big unaddressed problem with market data is the demand curve being somewhat inelastic. Every significant participant needs to subscribe to all feeds. Exchanges are free to foist fee enrichments with little recourse. If you want to be careful in fulfilling your best execution obligations in real-time, you need the data. There is not much choice. The rises, product differentiations, complex pricing schedules, and sometimes onerous volume levels serve little good. Even though the various SIPs' latencies have vastly improved, there is no substitute for direct feeds.

Due to such inelasticity and island monopoly behaviours, the regulator needs to regulate here. Market data is more a utility than a luxury. Some people may think, "If only we had a committee to look at this, we could call it something like the Equity Market Structure Advisory Committee." Personally, I think the SEC should just do its job. If it is not smart enough to do it, it should get smarter, but we know it is smart enough. It seems just a little less active than it should be. Perhaps the SEC is hampered by the new laissez-faire mood surrounding all regulators which may be being driven from the top.

Market data is a mess. Fix it, SEC.

HFT kangaroo loose in my top paddock?

Global warming has thawed some pundits' HFT frost. This is definitely one of the nicer things that made me go, "Hmmmm."

Frequent HFT critic and maverick Maverick, Mr Mark Cuban, acknowledged an investment in Virtu as a proxy for volatility. Though this has to be tempered with the idea that he was joining a firm he saw as advantaged rather than changing his mind too much about HFT. Bloomberg wrote in "Mark Cuban Once Trashed Speed Traders. Now He's Investing in One,"
That’s a turnaround from March 31, 2014, the day Michael Lewis’s book “Flash Boys” came out, in which he blasted high-speed traders. “There’s no such thing as bug-free software,” Cuban said that day in an interview on CNBC, calling it the greatest risk to that type of trading. “When you have fat-finger bugs you just don’t know what’s going to happen to the market.”
Bloomberg's Barry Ritholtz's change of stance was more of a change of stance when he reflected, "How committed are you to a belief system?" in the opinion piece, "One Question Investors Should Ask Now and Then - The answer might reveal flaws in your thought processes,"
High-frequency trading: I have been a pretty robust critic of HFT, based on the premise that trading is a zero-sum game. Whatever profits the HFT traders extract are coming out of slower-speed trading accounts. Sniffing out trades before they are executed and jumping ahead of them sure looks like front-running, which in the traditional sense is illegal. I don't know if this variant is illegal, but it sure seems wrong. 
My views on this have been moderated by (of all things) Vanguard Group Chairman Bill McNabb (interviews here and here). McNabb has made a fairly credible case that HFTs make it easier and cheaper for giant shops like his, which has about $5 trillion under management, to execute orders. It required a rethink of how the firm approached trading, but in the end it seems to be money-saver for a company that tries to pass on low costs to clients who are doing things like saving for retirement.
Even as Mr Ritholtz makes a concession, he includes a bit of a flawed thesis, on which the respected, Mr Adam Nunes commented via twitter,

Mr Nunes is so very right. This is a misconception I've been fighting since the newly appointed BT CEO, Mr Frank Newman suggested closing various IB businesses in the 1990s, including my prop trading group, Capital Markets Group (CMG). He suggested there was no point competing in zero-sum games. After the crowded lift failed, getting hot and sweaty as the minutes ticked by, I was wheeled out to choke a little infront for Mr Newman. I did spit out my prior approved question at the all-staff gathering to challenge his view on zero-sum. It was a young twenty-something to the slaughter. I still have nightmares about this wee little public choking and humiliation, but the point was made. Zero-sum begone.

Prior to the "Flash Boys" novella, I meandered about "markets are a zero sum game" being the biggest myth in all of finance, "Hedging – losses as profits that feed traders." Sometimes you have to count your wife's teeth.


Crypto-currency crime

Bubbles come and go. Meh. Being libertine at heart -  even though the libertine part seems to be shrinking with age - I'm disturbed by my disturbance at the marketing around so-called virtual currencies. Octogenarian family friends, who can't use an iPad, make enquiries about BitCoin. I suspect you, as a knowledgeable technology or finance peep, have been asked similar questions even if you are no longer reading this all too long meander.

If you are no longer reading I presume you will still know if you've been asked about such coinage - despite the breaking of the fourth wall. It is a worry that the National TV News from the public broadcaster here has had BitCoin prices and commentary in the nightly finance report for many months. Crazy stuff for something with the mirage of the modest market cap of a single stock. When remote Tasmanian retail, or the shoe-shiner, asks about Bitcoin, you know you have a retail problem that needs some unfortunate regulatory consideration.

I wrote a child-like meander about crypto-currency and crime last week, "Your crypto-currency response will be weighed." It serves little purpose other than as therapy but I enjoyed the music. The important aspect is to realise, according to the UN, there are some 40,000,000 modern slaves in the world and they deserve their AML back. There are kids being abused and photographed with their pictures being hosted on servers paid for with BitCoin.

Yep, slavery is still a thing:

My musing meandered me to the point of view that crime may destroy BitCoin before other factors, say bubbles. The regulators have been asleep at the wheel. Perhaps with its minor transaction flow profile virtual currencies didn't deserve too much attention, but the genie escaped that bottle years ago. The regulators could have nipped this nightmare in the bud with some pretty simple regulations. They failed their mission.

Most disagreed with this point of view in comments sent back, though I appreciated some support. I'm hoping this is another Søren Kierkegaard moment where my minority view may be right.

Society has a point. There is a minor fraction of real currency that is laundered. There is a minority, perhaps a much bigger minority of virtual currency that is likewise laundered. For the non-virtual case, we've developed some pretty flawed regulations and laws in a ham-fisted attempt to stem such criminal activities. They have a point as the best we can do for now. The same regulations should be equally applied to virtual currencies.

Money laundering is a pretty big deal,

(Sourced from PWC - click to enlarge)

unless you think $2 trillion a year is nothing to worry about.

I've been especially critical of the CFTC's and SEC's tepid response. That seems to be changing as they ramp up the rhetoric to counter their embarassing lack of action. See this sensible Written Testimony from the last week by Chairman J. Christopher Giancarlo. There is a lot of good work in there, including this piece on virtual currencies,
Let’s turn to virtual currencies. Emerging financial technologies are taking us into a new chapter of economic history. They are impacting trading, markets and the entire financial landscape with far ranging implications for capital formation and risk transfer. These emerging technologies include machine learning and artificial intelligence, algorithm-based trading, data analytics, “smart” contracts, and distributed ledger technologies. Over time, these technologies may come to challenge traditional market infrastructure. They are transforming the world around us, and it is no surprise that these technologies are having an equally transformative impact on U.S. capital and derivatives markets.
Supporters of virtual currencies see a technological solution to the age-old “double spend” problem – that has always driven the need for a trusted, central authority to ensure that an entity is capable of, and does, engage in a valid transaction. Traditionally, there has been a need for a trusted intermediary – for example a bank or other financial institution – to serve as a gatekeeper for transactions and many economic activities. Virtual currencies seek to replace the need for a central authority or intermediary with a decentralized, rules-based and open consensus mechanism. An array of thoughtful business, technology, academic, and policy leaders have extrapolated some of the possible impacts that derive from such an innovation, including how market participants conduct transactions, transfer ownership, and power peer-to-peer applications and economic systems.
Others, however, argue that this is all hype or technological alchemy and that the current interest in virtual currencies is overblown and resembles wishful thinking, a fever, even a mania. They have declared the 2017 heightened valuation of Bitcoin to be a bubble similar to the famous “Tulip Bubble” of the seventeenth century. They say that virtual currencies perform no socially useful function and, worse, can be used to evade laws or support illicit activity. Indeed, history has demonstrated to us time-and-again that bad actors will try to invoke the concept of innovation in order to perpetrate age-old fraudulent schemes on the public. Accordingly, some assert that virtual currencies should be banned, as some nations have done.
There is clearly no shortage of opinions on virtual currencies such as Bitcoin. In fact, virtual currencies may be all things to all people: for some, potential riches, the next big thing, a technological revolution, and an exorable value proposition; for others, a fraud, a new form of temptation and allure, and a way to separate the unsuspecting from their money. 
Perspective is critically important. As of the morning of February 12, the total value of all outstanding Bitcoin was about $149 billion based on a Bitcoin price of $8,800. The Bitcoin “market capitalization” is less than the stock market capitalization of a single “large cap” business, such as Disney around $156 billion. The total value of all outstanding virtual currencies was about $430 billion. Because virtual currencies like Bitcoin are sometimes considered to be comparable to gold as an investment vehicle, it is important to recognize that the total value of all the gold in the world is estimated by the World Gold Council to be about $8 trillion, which continues to dwarf the virtual currency market size. Clearly, the column inches of press attention to virtual currency far surpass its size and magnitude in today’s global economy. 
Yet, despite being a relatively small asset class, virtual currency presents complex challenges for regulators. Chairman Jay Clayton of the U.S. Securities and Exchange Commission (SEC) and I recently wrote: 
The CFTC and SEC, along with other federal and state regulators and criminal authorities, will continue to work together to bring transparency and integrity to these markets and, importantly, to deter and prosecute fraud and abuse. These markets are new, evolving and international. As such they require us to be nimble and forward-looking; coordinated with our state, federal and international colleagues; and engaged with important stakeholders, including Congress.
 It is this perspective that has guided our work at the CFTC on virtual currencies. Our work has six broad elements: (1) staff competency; (2) consumer education; (3) interagency cooperation; (4) exercise of authority; (5) strong enforcement; and, (6) heightened review of virtual currency product self-certifications.
The interagency mess is quite troubling. Self-certification is also troubling in this domain. We have the various SROs, SEC, CFTC, FinCEN, FBI, DOJ, FSOC, FSB, IOSCO all with a stake.

We have to be careful with our technology and regulations. Just because cameras are used for kiddie porn, we don't ban cameras. Similarly, mathematical truths and cryptography are too important for the good of humanity to limit or ban. The FBI and NSA are not too happy with solid cryptography being used in the wild. Shame on them. As a society, we have chosen another path. We choose to use society's lubricant, money, as one of the primary mechanisms to keep our children safe and to act against slavery. This, to me, remains a valid choice. Virtual currencies are subverting this due to a lack of oversight.

The docile and inactive FinCEN at least prosecuted the $4B laundry-o-mat that was BTC-e. A $110M fine was levied. Part of the case included highlighting not only the problem of bitcoin mixers but the clear and present danger of newer virtual currencies that provide untraceable features. Dash in this case:

With many next-generation coins, such as Zcash, and then Monero, improving on the usefulness of virtual currencies for crime, we need to wake up and process this nightmare that refuses to sleep. It is not possible for me to accept that Monero should be available for any reasonable transaction size given the impossibility of AML/KYC application to such a beast.

Australia introduced new virtual currency laws recently but a punter can still buy Monero at exchanges touting AML compliance. Many buy BTC at Coinbase and then convert to Monero.  It would seem there needs to be a central body or regulation that has viral properties. You can't expect global rules to be negotiated or exist - certainly not in a timely manner. A viral approach where you can't transact with a non-member or a non-compliant body is the only workable solution. A cabal of countries agreeing would be better, but the US or the EU could go it alone if necessary, again.

A further step would be for compliant venues to raise a Suspicious Transaction Report requirement for coinage going to non-compliant venues or being tumbled. It is right there in the chain and popular addresses are known or derivable. That is somewhat better than cash in that you can track the trackable, but not the new generation of crime-based coinage such as Monero. The viral possibilities in such an approach are better than what we have with cash. There is hope.

It does make you wonder if there is a grand conspiracy in that Bitcoin may be a long game where the US government is trying to corral all the criminals into a tight space where one sucker punch will bring many undone. Alas, government incompetence and Hanlon's razor makes this unlikely. Yet, it makes me go, "Hmmm."

Virtual currencies are a cesspool that need not be so. Regulators are belatedly picking up the ball, but not fast enough. ICOs and the vulnerability of poorly educated retail are one dimension but I worry more about crime. You should too. This is a thing that should make us all go, "Hmmm."


It was useful that the CAT provider, Thesys Tech, restructured to reduce conflicts. However, the CAT remains weirdly flawed. Not because of Thesys. The specification is just wrong. Not enough people care about this suggesting people are happy for it to fail in its mission even though it may succeed in its rollout. It's like saying a dingo is the perfect boomerang. CAT fees for nothing is a tiny bit disturbing. I doubt Mark Knopfler would be surprised:
Now look at them yo-yo's that's the way you do it
You play the guitar on the MTV
That ain't workin' that's the way you do it
Money for nothin' and chicks for free
Now that ain't workin' that's the way you do it
Lemme tell ya them guys ain't dumb
Dire Straits indeed. I maintain my old view on this described here, "Crazy CAT approved by SEC." It's absurd. This makes me go, "Hmmm."

Market volatility and value

One normal version of our parallel universe's VIX resuming escalated sedation:

What could possibly go wrong:

Valuations continue to make me go, "Hmmm."


It exists as a public exchange. That makes me go, "Hmmm."

Arms carrying arms

Since Sandy Hook, the USA has showed it can't fight its way out of a wet paper bag, or Congress, without a gun. That makes me go, "Hmmm." You need more data? Show who the fu*king data? Well, I guess you now have more fu*king data.

(click to enlarge)
The intention of the Second Amendment to do various things seemingly includes enabling state militia to overthrown a central government. How can they do that with a wimpy AR-15 with or without a bump-stock? Shouldn't cruise missiles, nukes, Mach 2+ fighters, stealth bombers, tanks with uranium tipped munitions, powerful explosives, rocket launchers, 50mm Gatlings, rail guns, patriot missiles, et cetera, all be available for all? Why the restrictions? How can the government get away with denying such obvious Second Amendment rights? How are you going to overthrow the Goliath that is the US government without powerful enough slingshots?

It's a haunting hypocrisy.

Australia used to have one to two mass shootings a year for a very long time. Roughly one every eighteen months. Then a retiring NSW realtor and his wife bought my parents' ten-acre hobby farm at Crabtree in the Huon Valley. Finally a peaceful retirement. The retired realtor's wife took his beemer and visiting guests down to the quite quiet and lovely Port Arthur for a scenic day out. His wife and guests were killed and one poor soul was bundled into the beemers boot as a coward killed thirty five people. The coward viciously hunted the very young daughters of a chemist. One daughter, Madeline, three years old, was lined up, point blank, against a tree as she tried to hide. Her mother was fortunate, in an awkward respect, to have not survived long enough to suffer the visions that haunt Madeline and Allanah's father.
John Oliver's series on gun control: Part One. Part Two. Part Three.
A gun buyback ensued. Twelve or so weeks of debate and legislation passed quickly even though a 'C'onservative government was holding court federally. How many mass shootings in the last twenty-two years, in total?


That should make us all go, "Hmmm."



No comments:

Post a comment