The retail investor now trades in milliseconds. The data telling us what that does to returns is finally arriving.

On 14 July 2026 a team of physicists published a paper that, for once, did not pretend the trading floor was the market. The study, indexed in Physical Review E and circulated by the Science X network on 14 July 2026 at 19:20 UTC, takes the decades-long migration of US equity trading from human specialists to electronic venues as a physical system and asks what it has done to the small investor in particular.

The shift, the authors argue, is not finished. Liquidity now sits in the matching engine; price discovery sits in the colocation cage; and the retail order, when it arrives, is filtered, routed and re-priced by intermediaries it will never see. A new study published this week gives the cleanest academic map yet of how that re-arrangement changed who gets filled, at what price, and on whose terms.

What the study actually measures

The paper combines two streams of data that are rarely stitched together: structural records of how US exchanges evolved from open-outcry and floor-based trading into fully electronic limit-order books between roughly 1990 and 2020, and trading behaviour data from millions of retail accounts over the same period. The authors treat order flow as a physical flow and apply statistical-physics tools to measure concentration, latency effects and path dependence in execution prices.

Three findings stand out. First, the average spread paid by retail orders narrowed dramatically across the period, but only after a clear inflection around 2005, when Regulation NMS in the US forced order routing through venues that quoted the best displayed price. Second, retail orders systematically received worse prices than institutional orders of the same size, a gap that persisted even after spreads compressed, suggesting the disadvantage had moved from explicit cost to implicit cost (price impact, latency, information leakage). Third, the introduction of payment-for-order-flow arrangements and zero-commission brokerage coincided with a measurable re-routing of retail orders toward wholesalers, with knock-on effects on the price at which those orders were filled.

The counter-narrative from the sell side

The dominant industry framing holds that electronic trading is the great democratiser. Commissions collapsed from tens of dollars per trade to zero; mobile apps put a US equities book in every pocket; the fraction of US adults owning stocks has climbed to multi-decade highs. None of that is in dispute.

What the industry framing downplays is the second-order effect. The retail order, once a social object executed by a known broker on a known exchange, is now a packet routed by smart-order routers across more than a dozen lit and dark venues, often terminating at a wholesaler that pays the broker for the right to handle it. The price the retail customer sees on the app is rarely the price the wholesaler prints on the tape. The new paper's contribution is to put a number on the size of that gap and to show it has structural, not cyclical, causes.

The standard rebuttal from brokers and wholesalers, articulated in comment letters to the SEC during the 2024-2025 payment-for-order-flow consultations, is that retail price improvement (the difference between the executed price and the prevailing national best bid or offer) more than offsets any implicit cost. The paper finds the opposite: in the subsamples it can isolate, retail price improvement is smaller than the implicit cost of routing, and the gap widens in volatile sessions when retail traders are most active.

What the structure looks like underneath

The deeper story is a familiar one in the economics of platforms. A market infrastructure that began as a public utility, the specialist system on the NYSE, was unbundled and privatised in stages: decimalisation in 2001, Regulation ATS in 1998, Regulation NMS in 2005, the eventual rise of dark pools and wholesalers. Each step was justified by a competition argument. Each step also shifted rents from a transparent venue to a stack of opaque intermediaries.

What the physicist's lens adds is the realisation that this is a phase transition, not a slow drift. Below a certain threshold of latency, the human trader is a curiosity. Above a different threshold, the retail order becomes a raw material for algorithms that quote around it. The retail investor does not trade against other retail investors. The retail investor trades against a market-maker who already knows the retail flow is coming, because the broker sold that knowledge upstream.

This is the structural point the paper makes without saying so in those words: the price the small investor sees is a retail price, and the retail price is, by construction, the price a wholesaler is willing to pay for the privilege of handling the order. The wholesale price, the one institutions transact on, lives in a different statistical universe.

What is still contested

Two questions remain open. The first is causality. The paper documents correlation between electronic-market structure and retail execution quality. It does not, and cannot, cleanly identify what would have happened to retail execution quality in a counterfactual market that remained human-centric. The retail-investing boom of 2020-2021, when zero-commission brokers and stimulus checks collided, is a confound that the dataset only partially controls for.

The second is policy. The SEC's 2024-2025 rulemaking on best execution, order competition and payment-for-order-flow produced a partial retreat from the status quo (additional disclosure obligations on wholesalers, a pilot for auctions of retail order flow), but the underlying architecture survived. The paper is agnostic on what regulators should do; the regulatory community has so far been agnostic on whether the paper's findings should change what they do. That is the conversation worth watching next.

Desk note: Monexus treats this as a markets and technology story, not a behaviour story. The wire service line emphasised retail participation and the post-pandemic boom in self-directed investing; this article re-centres the question on market microstructure, where the durable rents actually sit.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://www.sec.gov/rules/final/34-51808.htm
• https://www.sec.gov/news/press-release/2024-80