Most retail order flow routed to dark pools via PFOF actually receives *better* average execution quality than it would

Retail traders celebrating their "free" trades and fractional penny price improvement are unknowingly participating in one of the most sophisticated value extraction mechanisms in modern finance. The real cost isn't in your individual fills. It's in the systematic harvesting of your collective informational footprint, which quietly destroys alpha opportunities that independent quantitative traders would otherwise exploit from public market data.

The numbers seem paradoxical until you understand the business model. Citadel Securities and Virtu Financial internalize over 90% of retail marketable order flow in US equities, reporting average price improvement of $0.0034 per share according to S&P Global's 2023 data. Yet these firms simultaneously generate billions in annual revenue from this same flow. If price improvement were the full story, the math would not work. Someone is getting a bargain, and it is not the retail trader.

The prevailing narrative says otherwise. Wholesale market makers and their advocates point to SEC Rule 605 data showing that retail orders routed to internalizers receive better execution than the NBBO on over 90% of orders, compared to roughly 50% on exchange. A 2022 study commissioned by Citadel Securities estimated $3.6 billion in annual savings for retail investors through this mechanism. These numbers were cited relentlessly during the SEC's Reg NMS II deliberations. The logic feels airtight: measurable price improvement equals measurable benefit. Brokers market it. Industry funded research confirms it. And retail investors, conditioned to evaluate cost on a per trade basis, see no reason to question it.

But the price improvement narrative is dangerously incomplete because it treats each retail order as an isolated transaction. It is not. Aggregated retail order flow contains directional signals that internalizers can decompose, segment, and selectively trade against. The fraction of a penny you receive as price improvement is a loss leader. The product being sold is the statistical portrait of retail sentiment, timing, and concentration that wholesalers use to position ahead of predictable flow patterns.

This is where the mechanism becomes clear. Research by Battalio, Corwin, and Jennings in 2016, and subsequent work by Ernst, Sokobin, and Spatt in 2021, demonstrates that wholesalers systematically cream skim. They internalize the uninformed and profitable flow. They route toxic or informed orders back to lit exchanges. This creates an adverse selection asymmetry where lit venues bear disproportionate information costs while internalizers retain the clean, predictable, profitable side of the ledger. The price improvement they offer on retained flow is funded by the informational edge they extract from seeing the full distribution before deciding what to keep. Think of it as a below market rate purchase price for enormously valuable informational content. You are not receiving a subsidy. You are selling your data at a discount you never negotiated.

The second order consequences are where independent systematic traders pay the real price. When you remove retail order flow from lit venues en masse, you degrade the microstructural signals that quantitative researchers depend on. Order book depth asymmetries, arrival rate patterns, trade size clustering. These are the raw inputs for short horizon alpha generation. SEC Chair Gensler's 2021 and 2022 analyses showed that roughly 47% of all US equity trading volume now occurs off exchange, up from approximately 25% a decade prior. This is not a marginal shift. It is a structural transformation of the information ecosystem.

The academic evidence is consistent. O'Hara and Ye in 2011 and Comerton-Forde and Putniņš in 2015 demonstrate that as dark pool market share increases, lit market price discovery deteriorates. Bid ask spreads on lit venues widen by 4 to 8 basis points in affected names. Information share attributable to lit markets declines. And short horizon return predictability from order flow imbalance signals, the bread and butter of independent systematic strategies, degrades significantly. The edge has been transferred from a distributed ecosystem of quantitative researchers to a concentrated oligopoly of internalizers who monopolize the signal.

At Vhalanx Core, we do not simply observe this structural shift. We build analytical frameworks that account for it. By decomposing total market volume into lit venue, dark pool, and internalized components using FINRA ATS transparency data, Consolidated Audit Trail insights, and flow toxicity metrics like VPIN, we reconstruct synthetic order flow signals that model the missing retail component. This allows us to identify names where retail flow extraction has created the largest informational gaps on lit venues and to adjust our microstructure based alpha models for the systematic bias that dark pool segmentation introduces. The edge belongs to those who understand the topology of modern market structure well enough to correct for its distortions rather than naively consuming degraded public signals.

The SEC's proposed Rule 615 would require qualifying retail orders to be exposed to open competition via qualified auctions, potentially redirecting hundreds of billions in annual order flow back to price forming venues. If implemented, we may witness either a renaissance of microstructural alpha for independent systematic traders or an adaptation by internalizers so rapid that the informational extraction simply migrates to a new mechanism. The real debate is not whether PFOF helps or hurts retail on a per trade basis. It is who should own the informational exhaust of an entire generation's trading activity, and what kind of market ecosystem that ownership structure produces.

I am genuinely uncertain whether structural reform can outpace the financial engineering designed to circumvent it. I suspect the practitioners reading this are equally divided.