In late 2019, most major US retail brokerages eliminated commissions on stock and ETF trades. Retail investors, on average, celebrated the change as a clear win. The reality is more nuanced: the elimination of commissions did not eliminate the cost of trading; it moved the cost from a visible charge on the trade confirmation to an invisible cost embedded in the price at which the trade executed. Understanding where the cost went is worth every retail investor's attention.
Where the money used to go
Under the old commission model, a retail trade paid the broker a per-trade fee — typically $4.95 to $9.95 in the years before the elimination — as an explicit line item. The broker used this revenue to fund its operations, its research, and its profit margin. The trade itself was routed to whichever venue produced the best available execution, with the broker's incentive aligned to the client via the explicit fee.
The trade execution itself was not free even then. There was still a bid-ask spread — the difference between the highest price a buyer would pay and the lowest a seller would accept — and every trade crossed some portion of that spread. But the broker's revenue and the execution costs were separately visible.
Where the money goes now
The current model, standardised across major retail brokers in 2019–2020, is called payment for order flow (PFOF). Retail brokers direct client orders to specific wholesale market makers — Citadel Securities, Virtu, and a handful of others — which pay the broker a fraction of a cent per share for the right to execute the trade.
The market maker's incentive is to execute the trade at a price where they can capture a small profit relative to the current market. In liquid stocks with tight bid-ask spreads, this profit is genuinely small — often much smaller than a retail investor would notice. In less-liquid names, or in specific order types (particularly options), the price improvement or degradation can be more meaningful.
The regulatory structure requires that retail orders receive "price improvement" — execution at prices better than the National Best Bid and Offer (NBBO) — in aggregate. Data shows that retail orders through PFOF routing do receive price improvement on the vast majority of executions. The question is whether the improvement is as good as the improvement would have been under a different routing regime.
The empirical evidence
Academic studies of PFOF have found mixed results. Some find that retail investors receive genuine price improvement that outweighs any implicit cost, particularly in equity markets. Others find that the improvement is less than what would be achievable under alternative routing structures, with the difference accruing to the market makers rather than the investors.
The clearer finding is that the impact varies dramatically by order type. For simple market orders in liquid stocks, the difference between PFOF routing and alternative structures is typically small (fractions of a cent per share). For options trades, the difference is often much larger — options market makers under PFOF regularly capture substantial spreads that would not be available under different routing.
The retail cost is not zero
A retail investor placing a market order in a large-cap stock still pays the bid-ask spread. The spread is small — often one cent in a heavily traded name — but it is not nothing. Ten trades per day at a one-cent spread on 100 shares is $10 per day, or roughly $2,500 per year, in implicit trading costs that appear nowhere in the account statement.
A more active retail trader — hundreds of trades per year — pays materially more than the visible cost suggests. A less active trader (a few trades per year of buy-and-hold positions) pays a trivial amount. The elimination of commissions helped the less active trader marginally and helped the more active trader dramatically less than the headline "zero commission" implies.
The options case is worse
Options bid-ask spreads on retail-typical order sizes are often 5–10% of the option premium. On a $2 option, a 10-cent spread is 5%. On a $1 option, a 5-cent spread is 5%. This is an enormous cost compared to the underlying stock spread. It is largely a market structure issue rather than a PFOF issue specifically — options markets are less liquid than equity markets — but PFOF routing can amplify the impact for retail flows.
Institutional traders in options often achieve substantially better execution through algorithmic routing to multiple venues, price improvement auctions, and various forms of block trading. Retail options traders under PFOF do not typically receive the same benefits.
The behavioural effect of "free"
Beyond the direct cost question is a subtler behavioural effect. When trading appears free, people trade more. The Barber-Odean study and its many successors documented decades ago that trading frequency correlates negatively with retail investor returns. If eliminating commissions increases trading frequency — as it demonstrably did in the years immediately after 2019 — the aggregate effect on retail investor returns could be negative even if the per-trade cost is somewhat lower.
The 2020–2021 retail trading surge, with its notable overrepresentation of options activity, is a well-studied case. Aggregate retail returns during that period were mixed, and the concentration of losses in options-heavy accounts was documented in academic and regulatory analyses.
The rule to internalise
Zero commission is a misleading label. The cost of trading has not been eliminated; it has been moved from a visible line item to an invisible spread capture. For infrequent traders in liquid stocks, the change is a modest net benefit. For frequent traders and for options traders, the change is more complicated, and the "free" framing can encourage the trading frequency that historically correlates with worse retail returns. Reading the current structure clearly — rather than treating it as a customer-service upgrade — is the entry point to using it well.
Educational content only. Not investment advice.