How 40 Percent of US Stock Trades Now Happen in Dark Pools You Cannot See

The Iceberg Under the Surface

When you look at a stock trading at $50.00 on a screen you are seeing activity on the illuminated exchanges the NYSE Nasdaq and their affiliated venues where quotes and transactions are publicly displayed in real time. What you can't see is the approximately 35-45% of daily US stock volume that runs on alternative trading systems registered dark funds and broker internalization systems. This is documented in the company's off-exchange reporting data.SEC and FINRA that are published monthly. The existence of dark pools is not a conspiracy against retail investors. It is the rational economic consequence of having large institutional investors pension funds endowments mutual funds hedge funds who need to buy or sell positions worth tens or hundreds of millions of dollars without revealing their intentions to the market before finishing

The Three Main Types

Broker dark pools are operated by major investment banks Goldman Sachs runs SigmaType: Firms such as Citadel Securities and Virtu Financial receive retail order flow from brokers under pay-for-order-flow agreements and execute those orders internally. Citadel Securities alone handles approximately 40% of all US retail stock order flow. The third type is institutional matching networks such as Liquidnet which match large block orders typically $1 million or more between institutional investors anonymously

A Worked Example: What Hiding an Order Is Worth

The legend above states that a large illuminated order moves the price against you. That is the entire justification for this infrastructure so it deserves a number rather than a statement

Set up the trade. A pension fund needs to sell $200 million of a mid-cap stock trading at $50 that is 4 million shares. The stock trades an average of 2 million shares per day so its average daily dollar volume is about $100 million. Therefore the order is placed twice per full normal trading day

Estimate the impact on the market. The standard family of models used across the industry says that the impact scales roughly with the square root of the order size relative to daily volume multiplied by the stock's daily volatility. With daily volatility of about 2 percent and an order twice the average daily volume the impact is about 2 percent times the square root of 2 which is about 2.8 percent

If this is applied to $200 million the cost is approximately $5.7 million paid in the form of an average execution price worse than the $50 that appeared on the screen when the decision was made

If instead the fund finds a natural buyer on a cross-network and trades the entire block at the midpoint of the quoted spread it pays a commission and has essentially no impact. The savings with one order is around $5.5 million

That money does not belong to the pension fund operators. It belongs to the teachers police officers and firefighters whose retirement the fund must pay for. The reason this infrastructure was built and the reason regulators have allowed it is that the alternative costs beneficiaries real money on every major transaction

Now the other side of the ledger. Suppose the place is leaky. High-frequency participants detect the presence of a big seller after the first few runs and adjust their quotes before the remaining flow. The fund still avoids the full 2.8 percent but could return a third of the savings on worse investments in the unhedged part. Call it $1.8 million returned to someone else

That's the whole argument in a comparison. A clean dark run is worth about $5.5 million. A leaky one is worth about $3.7 million. Both outperform open trading which is why institutions continue to use these venues even when they suspect they are imperfect and why the quality of a specific pool deserves much more attention than the existence of pools in general

These are illustrative figures using a standard impact approximation rather than a calibrated model and actual costs vary wildly depending on stock urgency and execution strategy. The order of magnitude is the point

How the Trade Actually Prints

One thing worth clarifying because the word dark does a lot of misleading work. A dark pool deal is not a secret after the fact. It just gets dark before it happens

Each off-exchange run in US stocks is reported to a FINRA Trade Reporting Service and appears on the consolidated tape usually within seconds showing the price and size. It counts on the day's volume appears in the price history and is visible to anyone viewing the tape. FINRA also publishes weekly volume by individual alternative trading system so the market share of each named group is a matter of public record

What is retained is the pre-trade information: the quote the order the intent. A lit market displays a rest order before it is executed which is precisely what a large institution is trying to avoid and precisely what makes lit venues valuable for price discovery

So the accurate description is that the US stock market has full post-trade transparency and partial pre-trade transparency. That distinction is the basis of every serious argument on both sides of the debate below

The Market Quality Debate

It is actually debated whether dark pool activity improves or harms the overall quality of the market. The traditional argument in favor of dark pools is that they reduce institutional transaction costs which ultimately benefits end investors in mutual funds and pension funds. The counterargument is that dark pools fragment liquidity and reduce price discovery: if 40% of the volume is not visible on illuminated exchanges prices on those exchanges reflect only a part of actual supply and demand. The empirical evidence isMixed: Studies have found that trading in dark pools increases in periods of high volatility when institutions most want to avoid adverse selection suggesting that opacity plays a genuine role. But other studies have found evidence that high-frequency traders extract information from order flow patterns in some dark pools partially defeating the purpose

Case Study: Barclays LX and the Marketing That Was Not True

The concern about information leaks is not theoretical and the definitive demonstration came from a case brought by a state attorney general and not a federal market regulator

Barclays operated one of the largest dark pools in the United States called LX. Its pitch to institutional clients was that the site was safe from high-frequency predatory activity backed by a monitoring system the bank marketed as Liquidity Profiling which it said categorized participants based on how aggressive their trades were and restricted the most toxic ones

In June 2014 the New York Attorney General sued Barclays alleging that the bank had systematically misrepresented the composition of its pool. The complaint alleged that marketing materials shown to institutional clients understated the presence of high-frequency firms that a chart presented to clients eliminated a major high-frequency participant and that the Liquidity Profiles system did not restrict participants in the way clients had been told

The case was settled in early 2016. Barclays paid $70 million split between the state of New York and the Securities and Exchange Commission and admitted a statement of facts. Credit Suisse settled charges related to its own Crossfinder pool on the same day for approximately $84 million bringing the combined penalties to approximately $154 million the largest ever imposed on dark pool operators at the time

Two things emerge from that episode that point in opposite directions

The first is that concern about leaks was justified. The institutions that paid for protection against information leaks in these places did not receive what they paid for and it took litigation rather than market discipline to establish this

The second is that the failure was one of disclosure rather than concept. The problem was not that a broker managed a private venue. It was that the broker described the venue inaccurately to clients whose orders were there. This is a fraud problem that can be solved by an ordinary remedy and in fact was fixed. Anyone who uses this case to argue that dark trading should not exist is reaching a conclusion that is not supported by the facts

Where the Dark Pool Panic Is Overstated

The forty percent headline is used to suggest that most of the American stock market has gone underground.Four reasons why the reading is incorrect

Most off-exchange volume is retail and was never going to go on an exchange. A large proportion of 35 to 45 percent are retail orders internalized by wholesale market makers and those orders are usually executed at prices better than the public quote. Lumping them with institutional apple crossing under a scary number combines two completely different activities

Executions are protected by public listing. Under the NMS Regulation an off-exchange trade generally cannot be executed at a price worse than the best domestic bid shown on the illuminated markets. Public listing therefore governs private execution. This is why pricing has not collapsed despite decades of predictions: the dark places are valuing the enlightened rather than replacing them

Post-negotiation transparency is complete. As noted above each trade is printed to tape and volumes are published at the venue level by FINRA. Information asymmetry lasts for a few seconds not forever and there are markets in the world with much weaker disclosure that do not attract this attention

The real issue of price discovery concerns a much smaller number. If you exclude internalized retail trading the proportion of the volume that is genuinely institutional dark liquidity is well below forty percent. That residue is where the serious academic debate about whether electronic listings are being undermined really lives and it is a real question but it is not about forty percent of the market

My own opinion is that the structure is defensible and the specific locations deserve continued suspicion which is pretty much the opposite of how the topic is usually discussed

What It Means for Anyone Pursuing Finance

For retail investors the practical implication is modest: their order on a major broker is executed at a competitive price whether it is executed on an exchange or internalized. For institutional investors access to the dark pool is a significant determinant of execution quality;Moving large blocks without sending signals to the market is worth real money at scale. For anyone pursuing a career in institutional trading or market structure understanding the fragmentation of the US equity markets lighted exchanges dark pools internalization ATS and OTC is essential knowledge. The market you see on a screen is a subset of the market that actually exists

How I Actually Think About Market Structure

Market structure is the part of finance most often discussed by people with a position to defend so I try to maintain some habits that keep me out of the discussion and out of the evidence

The first is that I convert each claim into a cost. Saying a place is predatory means nothing until it is expressed in basis points versus the alternative. The worked example above is the format I use: how much the order costs here what it would cost there and who keeps the difference. Most of the debate about market structure evaporates when that translation is insisted upon

Second I read FINRA's weekly ATS data rather than comment on it because venue-level volumes are released and resolve a surprising number of arguments about who is actually trading where

Third I try to identify who pays and who gets paid in any deal. Payment for order flow exchange rebates and internalization economics all have a payer and after that money explains more why the market is this way than any argument about fairness

Fourth I take the Barclays case as a model for skepticism. The right question is should this place almost never exist. It's about whether this place does what it tells its customers it does and that is an auditable question with an auditable answer

That's how I approach it as a student trying to understand plumbing and not as someone with money on the line

The Bottom Line

About 35 to 45 percent of U.S. stock volume is executed outside of illuminated exchanges through dark pools of brokers like Sigma

Economics explains why it exists. A $200 million order on a stock trading $100 million a day has an estimated market impact of about 2.8 percent or about $5.7 million versus about $0.2 million for a block crossed at the midpoint. That $5.5 million difference belongs to the beneficiaries of the fund that placed the order

Barclays is the reason to remain skeptical about the individual locations rather than the concept. The bank marketed LX as protected from predatory flow the New York Attorney General filed a lawsuit in 2014 alleging that the marketing was false and Barclays and Credit Suisse paid about $154 million combined in 2016. The market you see on a screen is a subset of the market that actually exists and knowing which subset and who is getting paid the difference is the whole story.skill