Hello from the Dark Side

Vol #132: June 8th, 2023

The entrance to the floor of the NYSE is not through the towering Corinthian columns that adorn a cobblestoned Broad Street. 

The first obstacle is a tourist blocking white tent. After the voice on the walkie talkie crackles approval, you’ll drop down a few steps, and enter a low ceilinged foyer with reflective black glass. This is to remind you that the future of equities trading is really just a large server farm. Visitors can get a temporary badge here, and the clock punchers enter through another door further down the street. 

Metal detectors for everyone, with the suit jackets filing upstairs and the mesh jackets heading through the turnstiles for battle. Enter now into the halls where buyers meet sellers. If the pits look larger than life, the bell ringing podium does not. It’s an awkward three quarters height TV prop. 

Starting with the earliest days under the buttonwood tree, specialists and market makers have handled orders and matched interests at the Big Board. In addition to their uptown rivals at NASDAQ, there are also regional exchanges from the Pacific to Cincinnati that offer some form of equity trading. The Chicago Stock Exchange even has a little trading floor perched above the off ramp to I-90.

Created with Midjourney

Exchanges are not the only game in town however, you can trade your shares elsewhere - almost 40% do. Long before there were decentralized protocols running on blades, dark pools were monitorless CPUs daisy chained together, carving out increasingly important percentages of equity market share.

A relatively benign regulatory change called 19c3 passed in 1979, allowing for off-exchange trading. Not much happened for almost a decade though, as practically an exchange floor was still the best place to meet and transact. Anyone who was going to give you a decent market on your General Electric stock was going to be there anyways. 

What Nvidia chips are doing for AI, IBM computers did for 1980s equity markets. Instinet launched the first dark pool in 1986 to allow for after hours trading. The next year ITG opened an intraday pool called “POSIT”  and traders have been doing it in the dark ever since. 

Dark pool trades are OTC or over-the-counter trades. That broadly means that two sophisticated actors agreed to exchange a certain quantity of stock for a specified price. These licensed broker dealers are then required to notify the world at the aptly named “Trade Reporting Facility” (TRF) within ninety seconds. 

The trade will hit the consolidated feed and get officially disseminated to the tape watchers, but it typically has much less meta-data than an exchange trade, in addition to being significantly delayed. A minute and a half is an eternity in today’s electronic market. HFTs are shaving picoseconds by soldering algorithmic logic directly into microchips; a trade report from that long ago might as well have come via the Pony Express. 

A delayed trade report that doesn’t give much information to observers, might be exactly what the transactors want. If you’re buying or selling large quantities of stock, the biggest problem you have is slippage. Liquidity imbalances mean prices shift, and it’s not in favor of the market taker. Electronic markets exacerbate this temporarily, with nimble market makers rapidly adjusting their pricing and sizing.

Because they run such an adverse selection bias, and have to keep large amounts of capital constantly at risk if they want to trade at all, electronic liquidity providers can be very skittish. Every buy or sell order is treated as the tip of an iceberg until its toxicity can be appropriately dissected. 

Dark pools solve for this nicely. Not only does your big trade not hit the consolidated feed until it’s too late for anyone to over react, but it can sit there hiding in the dark, patiently nibbling away at contras. 

Most professionals are reading “Level II” market data. This is one drill down from your basic size and price information. Level II shows you the market depth, or where there is interest at the various prices above the best offer or below the best bid. 

Courtesy of Bulls on Wall Street

If you’re a human watching this, and you see a new line pop up on the left hand side that says 192.45  followed by a 99, it’s time to fade your offers, there’s a big buyer who just slammed open the door. If you have shares to sell, let’s wait until Mr. Bid has satisfied his appetite. Computers are watching and processing their adjustments near instantaneously. 

The Level II montage only shows public “lit” exchanges; Mr. Bid can quietly post that order on any one of a dozen dark pools and no one is the wiser. Eventually savvy data crunchers will realize in real time that there’s some density in the order book and navigate accordingly, but the order is never officially announced until it has been traded. 

Dark pools also offer the potential for creative order types where public exchanges lack the regulatory flexibility to deliver. One of the more common is midpoint matching. This truce between buyers and sellers prints trades halfway between the bid and offer of the lit markets.  No one has to do the undignified and cross a spread. 

The irony here is apparent. Dark pool traders don’t want to reveal themselves in the light, but they need other people to actively trade there so they can get a fair price. It’s not dissimilar to the fear of too much passive investment. If everyone’s on an index auto-buy program, or no one posts in lit markets, what sets the price?

I’ve worked with traders who swore by the efficiency of executing their stock OTC, and others who didn’t know what a dark pool was. I never saw anything statistically significant about execution algorithms that used them exclusively, but I still appreciate why you might choose a dark only strategy. 

Importantly, dark pools only truly exist for equities. Options have no trade reporting facility. Even if a broker has matched buyer and seller through phone, IM, or smoke signals, the order must cross on an exchange. Equity trades get reported, options trades get exposed. 

Market participants almost always have the ability to break up an options cross by providing a better price. While the equity cross has to be within certain price bands, there’s never any risk of either side getting cut out. Option orders get broken up and price improved all day. While that’s usually good for the customer, there have been enough arguments foisted that allow for special carve outs. 

Qualified Contingent Cross orders can trade without break ups if they meet size and hedging requirements. Facilitation and Solicitation orders have their own nuances about degree of price improvement and minimum size. FLEX trades almost always go up unimpeded because of the short window to evaluate a non-standard trade structure. 

The closest the options market has to a dark pool is a sandwiched concept we called a pre-exchange mechanism. Because every order had to be crossed at an official exchange, the rules of engagement for this electronic box only determined who would go on to the final round. 

Electronic paper from retail brokerages is the golden goose of orderflow. The right to trade this is heavily bid up by deep pocketed interests via payment for orderflow. (Which I’ve argued is generally good thing.) To prevent the oligopoly from shrinking further, these trades are randomly parsed out on a wheel to the various competing interests. 

Every month the stats are evaluated for price improvement and fill rates, and the relative proportions of the pie might shift. Brokers are careful never to let one slice get too big, lest they lose their own pricing power. In order to get on, and stay on the wheel, you need to be able to demonstrate consistently high performance across the entire market. 

While the biggest shops trade everything from AA to ZVZZT, there are top tier players that choose to specialize. Highly liquid index and ETF products might work well for a firm with an HFT bent, while lower market cap and more speculative flow fits a mid frequency strategy. What if these frenemies banded together to divide up a common slice of the pie?

Better yet, with a little bit of competition, perhaps this collective could generate better pricing than any one single entity. By creating a pre-exchange mechanism where firms bid against each other to win the right to cross an order, the operator of this pool gets a shot at handling a growing share of the options market pie.

Winning the pre-exchange mechanism meant you earned the right to submit the customer order to a public exchange price improvement auction. As the initiator of the auction you’re conferred certain rights of participation, but the ultimate trade price is still determined by open competition. It’s not exactly a dark pool, because the order eventually trades on the exchange and shows up in your OPRA trade feed. 

It’s up for debate whether a true dark pool would be a good or bad thing for options. One major fear is that it would further consolidate the number of liquidity providers. If wholesalers can bypass the exchange and trade directly with customers (via their broker) then the best connections will quickly monopolize those pipelines. This reduces on screen liquidity and makes it more expensive for everyone to trade. 

Truth be told though, we haven’t seen that in equities markets. Certainly a cottage industry of execution algorithms has boomed as market structure complexity has metastasized, but overall liquidity has also grown. 

A major reason to keep options liquidity relatively centralized (yes I know there are 16 exchanges) is that they are significantly more complex. By my count there were 664,361 strikes listed at yesterday's close, or an average of 113 per product. With calls and puts that's over 200 times the number of bids and offers needed for an equity.

If your objective is to trade quietly and stealthily, there are many alternatives. From a technical approach, better iceberging algorithms are constantly being developed. If you can find the right rolodex, there are only a few doors you need opened to get one stop liquidity that doesn’t leak on the street, no matter how big your size. 

Perhaps the biggest myth of dark pools is that they’re not really dark. Someone is always watching, trying to reverse engineer what’s inside. Through pings, indications of interest, and closely monitored fill rates, data monsters are building an ever evolving map of the liquidity landscape. Your order can only hide for so long.

Created with imgflip.com