That Pool is Darker than You Think

An exploration of the outcomes and consequences of “dark pool” trading.

Market makers and broker dealers trade most shares of most companies in “dark pools”, which are “non-lit exchanges”, that is, a way to trade and settle without price discovery. The argument is that this is necessary for block trades (large purchases or sales of major institutional investors who wants to maintain the existing share price despite their transaction activity, which is by itself arguably a violation of the core principles of markets, but that’s a discussion for another day), but block trades are rare for any specific stock or trade on any specific day. They do happen, but infrequently, and certainly not to the extent that the rationale for dark pools can be justified by the block trade excuse.

And that’s what the block trade rationale is, an excuse.

So why do market makers and broker dealers trade the predominance of shares in the predominance of companies in dark pools instead of lit exchanges?

At some point since the digitization of trading, especially the adoption of high-frequency trading, market makers and broker dealers have realized that dark pools give them near total control over the price activity of the markets.

Market makers and broker dealers are not honest, straightforward servicers that act upon and for the best interests of their “clients”, the actual traders that make use of the exchanges. In fact, with payment for order flow (PFOF), the largest “client” of a market-maker or broker-dealer is actually themselves. The market-maker arbitrages the spread, while the broker-dealer loans out their “clients” shares to short-sellers who need to settle (or pretend to settle) obligations resulting from their short-sales. So neither the market-maker or broker-dealer has a fundamental incentive to act in the best interest of the traders, since they both explicitly benefit from the trading activity regardless of the harm inflicted on the traders.

But the more harm inflicted, the more the benefit available to the conspiring parties, so the incentive demands that over time, in maximizing the revenue and profit-making activity of self-dealing, necessarily the harm to the nominal “client” (traders) is increased.

What was originally tolerable as a nearly-square deal, not entirely on the up-and-up but good enough not to bother arguing over it, where the MM and BD nibbled at the edges of trader activities, due to the typical competitive forces of everyone in a market adopting the same winning tactics, then maximizing the control space of those tactics, once MMs and BDs realized the power of dark pools, the core incentive was to worsen the deal further and further for traders to the benefit of MMs and BDs.

How does the MM or BD benefit from dark pools? It’s quite simple, actually.

There are two directions a stock can go – up or down. The MM or BD has its own book of activity that benefits, on a case-by-case basis, from certain trades trending a share price up or down. More specifically, the MM will have a “short book” of short sales made (purportedly to “create liquidity” and ensure an active market regardless of actual market activity) that will (supposedly) be purchased at a later date (securities sold but not purchased) and a “long book”, which is what the average person typically thinks of as a stock portfolio.

When a MM/BD has a long position on a stock, they make purchases of those shares on the lit markets, and sales of those shares on the dark pool – that way, to the greatest extent possible for the malicious actor, only buying pressure exists on the shares.

When a MM/BD has a short position on a stock, they do the opposite – the make sales in the lit markets, and buys in the dark pools. That way only the selling pressure exists on the shares.

But wait! Every purchase is a sale, right? Since someone is selling and someone is buying? Not exactly. The ticker moves green (up/buy) when the bid is matched above the ask, and the ticker moves red (down/sale) when the bid is below the ask. When they happen to exactly match, the ticker is grey and there’s no price action – there’s also no arbitrage, since there’s no price difference, so there’s no way to make money on a neutral trade (I think there is, but we’ll get to that later, for now we can disregard it).

And what PFOF does is give the MM/BD enough prior notice of the trading action that it can perform that match-making to steer the price action, and once the matching is done, steer the actual trade activity to the lit markets or a dark pool, as may be most beneficial to the malicious actor at that specific moment in time.

Fundamentally, what internalization of trading and PFOF does is provide a dichotomy in the market that selectively enables malicious actors to steer buy and sell pressure in a direction that benefits the MM/BD, while actively harming the traders and the company that enable (and subsidize!) these malicious activities.

In short, a MM/BD makes lit purchases and dark sales for their long book, and dark purchases and lit sales for their short book.

But, like a classic discovery process, unfaithful-lover-trickle-truthing, or classic gameshow host, that’s not all!

The Contract for Difference Illusion

There’s also contract-for-difference, or CFD. CFD is a legitimate practice, when disclosed to a counterparty that understands what’s happening. Based on the Gamestop, AMC, BBBY, and many other current and historic company “cellar boxing” attempts (and successes), as well as the naked fraud recently exposed among many centralized crypto “exchanges” (not your wallet, not your money, calling them an “exchange” is a sweet lie to disguise what really happens in CEX, which is that, as the name implies, you get fucked, but not in a good way), I believe that most BDs are engaging, at least on some level, in CFD “trading”, where they never actually purchase shares, but instead “promise” to purchase a share on behalf of a traders order, credit the account for the value of the “purchase”, and then use their collusion to steer share prices.

Much like FTX, the “broker” takes your money, credits the shares you “purchased” to your account, then uses their price steering powers to ensure that they can not only spend your entire payment amount however they want (since they don’t have to actually purchase the security they claim to purchase), but they can do their best to maximize that profitability when you decide to exit the trade eventually too – a classic double-dip.

This CFD-style activity further increases the ability of the MM/BD to separate their buy-book from their sell-book to neuter price discovery, as any trading action of a client that the MM/BD doesn’t want to create price pressure can be managed as a CFD, even if only temporarily until a good match can be found – more on that later – and enables significantly more selectivity about when buy and sell pressures actually realize themselves in the market, even absent any dark pool obfuscation.

It seems entirely possible, plausible, and perhaps even overwhelmingly likely, that many classic broker-dealers are actually CFD platforms. We find more evidence of this from the behavior of the Jan ’21 “sneeze”, where BDs “shut off the buy button” on a list of securities. If the BDs were actually buying and selling shares, and the MMs were actually match-making, there’s no conceivable means that legitimately purchasing and selling shares on behalf of clients could cause a problem for anyone.

But if those BDs and MMs were actually using trade internalization to manipulate prices on their own long and short books, and BDs were pretending to purchase and sell shares when they were actually leveraging their price manipulation to guarantee themselves profits on a CFD “phantom” trading activity, that explains why there was so much widespread panic on Jan 21 – the tide went out unexpectedly, and everyone that’s been fucking the public for the last 100 years was left completely naked and exposed, as they deserve. When the price shot up so quickly, the CFD calculations were completely underwater to the extent that the MM/BDs no longer had sufficient margin on all the leverage they’d taken on, and the bank margin calls started – like with Robinhood that had to scare up a few billion dollars in a few hours. How would they be that short, that fast, if they were actually buying and selling the shares they said they were? The answer is, they couldn’t be, if they were trading legitimately as advertised.

In my opinion, “shutting off the buy button” is strong, if not overwhelming, evidence that a significant number of BDs are actually engaged in CFD and not actual trading. Paper trading, but taking your actual money, then making sure that you lose as much as possible. A casino game disguised as an investment run by a scam artist disguised as a financial services company. Oh, sorry, that last one is a big redundant, isn’t it? Combine the buy-button issue with the transparent CFD behaviors that recently collapsed crypto “exchanges” have recently exposed and we have a very compelling case for what caused the “sneeze”, and why the involved parties behaved the way they did.

But again, that’s not all!

ETF Infinity Shares

Crime is as old as humanity, and everyone on Earth has engaged in it at some point. There’s no one innocent, so let’s put that aside. New crimes, new criminals, and new innovations in dynamic criminality emerge every single day, and will never stop. One enormous new tool in the criminals toolkit are ETFs, or exchange traded funds.

ETFs have a lot of benefits, absolutely. So do swords and carbon dioxide. But, also like swords and carbon dioxide, a good thing can be used for terrible purposes in the wrong hands. And unfortunately, most of the people who invent new financial tools don’t do it because they have good hearts and love their communities.

A classic ETF has multiple securities in it, so that the securities can be traded as a category. This makes it easy to invest in a specific market in general, without having to suss out which companies in that market are the best bets, or just to limit exposure and volatility, particularly in emerging markets where there may not be a clear winner yet, or you just want to benefit from growth in the industry as a whole. All good things! (New ETFs can include just one security, which is a whole other ball of vomit and cheating I won’t get into here.)

A few characteristics of the ETFs are the fungibility of the baskets, the cash equivalency of the shares in the basket, and the need to regularly “rebalance” the ETF against the independent market activity of each of the securities included in the ETF.

The basket is fungible in the sense that no single ETF share can be discerned from another, in the sense that the ETF isn’t “real” – there’s no shareholder certificates registered for an ETF at the DTCC, because the ETF doesn’t actually exist as a going concern, it’s a product of MM/BD “innovation” (for better or worse). Nobody can tell the difference between this ETF trade or that ETF trade, and the ETF can’t be direct-registered.

An ETF is an assemblage of underlying shares, each with its own pricing, and rolled up into an ETF whose price is supposed to, theoretically, reflect the prices of the underlying shares. But remember, these are fungible, so they can’t be ascribed to a specific trade or a specific shareholder certificate. What that means is that ETFs can be settled on a cash equivalent basis. If someone buys $50k of ETF shares, and the “seller” doesn’t have the actual ETF shares to settle the trade, they can provide a “cash equivalent” of the number of shares of an underlying security, and that is sufficient to settle the ETF obligation from the seller to the buyer.

This cash equivalency means that shares are constantly leaking out of, and into, an ETF through the cash-equivalent settlement activity.

Now let’s pretend, just for fun, that you have a share you’re short on, and other shares you’re long on, and they’re all in the same ETF. And someone buys shares that you’re short on, so you can’t settle with those shares. But you can send them shares you’re long on, as a cash-equivalent, and the BD will settle that trade as an ETF even though they just received shares from one (or more, any mixture really, as long as it’s cash-equivalent) different companies than the actual distribution of shares in the ETF.

That implies that as long as the MM/BD maintains a position on any ETF that a targeted company is listed in, it can use the cash equivalent of any other listed shares to settle trades on that ETF!

And these shares can not only be “combined” through cash equivalency, they can be “broken out” that way too! You’re short on a company and can’t get its shares to settle an obligation? Aw shucks, that sounds like a pile of trouble! Why not take shares you’re long on, that are in an ETF with the shares you need, trade that ETF “basket” for the cash equivalent, take that cash equivalent as the shares you need (the ones you’re short on), and viola! Now locating and delivering the shares you committed but didn’t have is magically someone else’s problem! The next guy in line who thought they were buying a neutral ETF, is now stuck for the shares you were pretending to deliver. And the hot potato goes around and around, and where it stops, nobody knows, because it’s all fungible!

The next guy is short on those shares, but remember it’s all cash-equivalent, so they just buy a new ETF cash-equivalent, and boom, problem is solved – for them. Not for the buyer who is still waiting on the shares they purchased that have been failed to deliver, thus far, but you can roll that one over indefinitely, as long as you can use cash-equivalency to settle. Nobody ever has to actually engage in price discovery, nobody has to actually locate shares of the actual stock being purchased, people can just keep playing profitable games with ETFs until the sun burns out.

But remember that the ETFs aren’t the actual shares, attached and obligated, they’re just a representation of those shares. The ETF needs to be rebalanced on occasion to true up the share counts listed as committed and the pricing of those shares to ensure that the ETF reasonably matches the count and cost of the underlying securities – because remember, there’s no actual attached securities, so there’s no direct physical (well, metaphysical) connection between the ETF and the shares it represents, so the two will drift over time.

That drift over time that results from the fungible baskets and cash equivalency with on underlying shares forces the need to rebalance the ETF. But imagine you happen to be in a position where you know what the share prices of the underlying shares are, how available each of the shares in the ETF are in their “natural” (non-ETF) form, and can identify the difference between the ETF’s representative value as a basket, and on a per-share basis, and the actual current trading price and availability of those shares.

Golly, it seems like the drift between ETF pricing that requires regular rebalancing also creates an interesting arbitrage opportunity between not only the delta between ETF-cash-equivalent and actual-trading-price based on the basket distribution, but also an interesting arbitrage between the drift on individual shares! Imagine if, perhaps, someone was long on some shares in an ETF and short on others, and they could monitor that drift between ETF and actual shares, and then engage in cash-equivalent trading behaviors that would maximize the difference between the ETF and the actual share prices to their advantage! Now we’re arbitraging not only the ETF itself, but arbitraging the relative drift between different shares in an ETF as the trading patterns of the constituent shares send some prices upwards and others downwards.

This creates an opportunity for a malicious party (if one existed, surely they don’t, since criminal behavior is uncommon among humans) to leverage that price drift information to maximize the differential between different constituent members of an ETF, and purchase shares whose ETF representation is underpriced in order to fulfill orders for shares whose ETF representation is overpriced! So you get an order for say, AME, which you’re short on, and you definitely don’t want buying pressure on that one, and it happens to be in a ETF with, let’s call it GZN, which you’re long on, and definitely want buying pressure for. You purchase the cash-equivalent of GZN, call it the ETF, send yourself that ETF (or a complicit counterparty), and they “break it out” into the cash-equivalent shares of AME. Boom! You just took short-book buy-pressure, completely neutered it by turning it into a long-book buy pressure, and also increased your margin by arbitraging the difference between the underlying share prices and the ETF drift.

At this point, it’s crimes on top of crimes on top of crimes on top of more crimes, and everyone turns a blind eye, because the only people who it hurts are the people nobody cares about – that is, people not actively engaged in criminal activity in the financial markets.

Time Arbitraging Settlement

And what if you’re doing this and you realize that despite computer systems discovering, matching, and potentially settling trades on the nano-second timeframe, the actual obligation to settle doesn’t occur for two entire days? This introduces another interesting quirk – you don’t have to take the first match you get, as one might assume. In fact, you can pretend to take a match, claim to the involved parties that the deal is done, but actually sit on the trade for the 2-day period.

And I bet, with enough volatility, or fast enough trading (maybe, say, a higher-frequency style of trading than normal people can use?) within that two-day period, if there’s enough trading activity, you can probably find an even better spread between the buy and sell parties. When this is possible, it actually doesn’t make sense to complete matches immediately, even if it needs to appear that way, because the best result would come from holding all the trades and only finishing the match to perform settlement when the timeframe is over, all possible trades within the timeframe have been reviewed, and the best possible match can be made that uses the time between actual matchmaking speed, and obligate regulated matchmaking timeframes, to find the absolute most profitable matching for the trade.

Instead of the buy-book and sell-book being matched in real time, to the best of anyone’s ability, time arbitraging the trades completely separates the buy-book from the sell-book, severs any direct relationship between those activities, enables the buy-side and sell-side to be steered to the greatest benefit of the responsible party, enables the use of ETF tricks to create supply and demand as needed to balance any inequalities between the books, and gives everyone involved several days to find the best possible mélange of these manipulations to maximize their revenue and margin at every step.

This indicates that time arbitrage for settlement works for internalization, for CFD, and for ETFs, so you can actually take all these different forms of crime and stack them on top of each other for more and more and more profit, with essentially zero risk. And who’s going to stop you? 

Whew! Are we having fun yet? Because this shit is wild. If any of this speculation and conjecture is true… well… let’s just hope it’s not, because I can only imagine the consequences if this kind of malarky was going on out there in the squeaky-clean and definitely-never-criminal financial services industry.

Hiding Internalization through ETFs

ETFs provide another nifty feature in that they can hide internalization to further distort the use of dark pools to conceal trading activity, simply through the MM/BD making use of the ETF fungibility and cash-equivalency. Let’s say that the MM/BD has a ton of trades that they can’t match during the time window they can arbitrage to get the best pricing. Well shoot, find an ETF that the shares are included in, and just settle the trade by buying the ETF shares (or selling, depending on the trade direction) and then breaking out the ETF into whatever constituent shares you need.

Once MM/BDs realized that they had all these new tools at their disposal to completely control price discovery, match making, and trading patterns, while almost entirely concealing their activities through ETFs, dark pools, CFD fakery, and time arbitrage. With so many tools, the trading patterns can be shifted around in real time, as necessary, and make use of innovative ways to create margin at every step, no matter what direction the MM/BD needed to shift on a moment-by-moment basis.

This flood of fraudulent liquidity leads to a pretty dramatic Walter White situation – what do you do with all that cash!? Where you can you stash it?

Especially when any investment of this magnitude needs to be hedged, because the people the cash is obligate to are not exactly kind and forgiving persons. So how can someone with billions of illicit gains hide that money when there’s only so many Rembrandts and Monets in the world?