Payment for Order Flow: What It Means for the Price You Actually Get

You clicked buy on 200 shares at 9:41 on a Tuesday. The screen showed $47.02 bid, $47.04 ask. The fill came back at $47.033, no commission, and the trade confirmation had a blank where a fee used to be.

You still paid something. Against the ask you saved $1.40, because 0.7 cents a share on 200 shares is real money coming back to you. Against the midpoint of $47.03 — the price that sits halfway between what buyers were bidding and what sellers were asking — you paid 0.3 cents a share, or $0.60. Payment for order flow is why your broker was glad to route that order to a wholesaler instead of an exchange. Price improvement is why the fill beat the posted ask. Both are true on the same trade, and most explanations pick one and ignore the other.

Where the 200 shares actually went

Your broker did not necessarily send that order to an exchange. In a typical retail arrangement it goes to a wholesale market maker, which pays the broker a fraction of a cent per share for the right to handle it. The wholesaler fills you from its own inventory, at a price inside the posted quote.

Why pay for the privilege of trading against you? Because retail orders arrive in a steady, uninformed stream. A market maker buying from a hundred small sellers and selling to a hundred small buyers captures part of the spread over and over without getting picked off by someone acting on information it lacks. That is the whole business. FINRA’s investor overview of stocks lays out the bid, the ask, and the market maker’s role from the regulator’s own point of view, which is the cleanest place to start if the vocabulary is new.

The benchmark everyone reports against is the National Best Bid and Offer: the highest bid and lowest offer across every venue quoting the stock at a given instant. Your broker owes you best execution, and the practical yardstick is whether your fill was at, better than, or worse than the NBBO at the moment the order arrived.

That yardstick has a weakness worth naming once. It is a snapshot. Your order arrived when the quote read $47.02 by $47.04. If the market had drifted to $47.01 by $47.03 four hundred milliseconds later, your $47.033 fill still gets logged as price improvement against the first quote, even though it looks poor against the second.

The number to carry in your head

Half the spread, divided by the price. That is the whole cost model for a marketable order, and everything below is that one fraction applied to different situations.

For the 200-share trade: the spread was 2 cents, half of that is 1 cent, and 1 cent on a $47.03 stock is 0.0213%. On the $9,406 position that is $2.00 if you had filled at the full ask. You did better than that, but $2.00 is the ceiling.

Notice what the percentage does not depend on. Share count cancels out entirely. A 1-cent spread costs 0.0106% of your money whether you buy 10 shares or 5,000. A 25-cent spread costs 0.266% at either size. Order size changes the dollars, never the percentage — which is why the table below is useful for budgeting dollars and useless for deciding whether a trade is “too small to bother with.”

Every figure assumes a $47.03 midpoint, a marketable buy filling at the ask in one clean print, no commission, and no regulatory fees. These are hand-computed illustrations, not measured fills from anyone’s broker.

A sceptic reading that bottom row is right to object. A 5,000-share order does not get one clean print at the ask; it eats through the displayed size and the rest fills higher. The table is a floor, not a forecast, and it understates the cost badly in anything thinly traded.

What price improvement gives back, and what it doesn’t

Your fill beat the ask by 0.7 cents. The half-spread was 1 cent. So roughly 70% of the half-spread came back to you, and the market maker kept 0.3 cents a share on your side of the trade — $0.60 — plus whatever it captured on the offsetting side when it sold to someone else or bought from a seller a moment later. Your broker collected its per-share payment separately. Nobody in that chain worked for free, and none of it appeared on your statement.

Now compare the two levers you have. Push the improvement rate from 0% to 80% on that 2-cent stock and you save 0.8 cents a share, which is $1.60 on 200 shares. Trade a stock with a 1-cent spread instead of a 20-cent one, holding the improvement rate constant, and you save 9.5 cents a share, or $19.00 on the same 200 shares.

Spread width beats routing by roughly an order of magnitude. Which stock you chose to trade matters far more than which venue handled the order — and that is the practical conclusion buried under most arguments about payment for order flow.

Multiply by how often you do it

One trade at $0.60 over the midpoint is noise. The question is what happens across a year.

Take a $10,000 position, bought and sold in full each round trip, so the spread gets crossed twice. Assume a 2-cent spread with 50% price improvement, giving a net cost of half a cent a share each way. At $47.03, $10,000 buys 212 shares. Each crossing costs 212 × $0.005 = $1.06. A round trip costs $2.12.

Run it backwards and it gets more useful. Suppose you want total spread cost held under 0.25% of the position per year. That is $25.00, and at $2.12 a round trip it buys you eleven round trips. Trade a stock with a 5-cent spread instead and each round trip costs 212 × $0.0125 × 2 = $5.30, so the same budget funds four round trips. Same account, same rule, a quarter of the activity.

At the other end: hold that position three years and the one-time $2.12 amortises to about $0.71 a year, or 0.007%. Arguing about routing on a three-year hold is arguing about the wrong thing.

Zero commission is not zero cost, but it isn’t a con either

The cost moved. It used to be a $6.95 line item; now it lives inside the fill price. Whether that helped you depends entirely on your size.

On the 200-share order, the half-spread cost $2.00 at worst and the old commission would have been $6.95 on top of a spread you were paying anyway. The flat fee was seven times the spread cost and it did not scale — small orders subsidised the pricing model. On a 5,000-share order at $235,150, that same $6.95 is 0.003% of the position while the 2-cent half-spread is $50.00, or 0.021%. The commission all but disappears.

So the honest version: for small orders the commission dominated and its removal was a genuine saving. For large orders it was already a rounding error and its removal changed almost nothing. What did not change in either case is that you cross the spread every time you trade.

Options are where the percentages get ugly

Payment for order flow is largest per unit in options, and the arithmetic explains why. A contract quoted $1.20 bid, $1.30 ask has a $1.25 midpoint and a half-spread of 5 cents. Multiply by the 100-share contract multiplier and that is $5.00 per contract against a $125 position value — 4% of the trade, on one side.

The same 10-cent spread on a $47.03 stock costs 0.106%. Roughly a thirty-seven-fold difference in percentage terms for an identical dollar spread, entirely because the contract price is small relative to the tick. Do not carry the equity numbers over.

When these figures are wrong

Your order is large relative to displayed size. If 300 shares show at the ask and you want 2,000, the balance fills higher. On a stock holding a few hundred shares per price level, that order might average 4 or 5 cents above the touch rather than 1 cent — four or five times the table.

You traded in the first or last minutes of the session. Spreads at 9:30:05 routinely run several times their midday width. A name quoting 2 cents at 11am might quote 12 cents at the bell. Use the 10-cent column.

News just hit. Market makers widen quotes when they cannot tell whether you know something they don’t. The half-spread cost during that window can exceed a whole year’s worth of the turnover table.

You rested a limit order and it got filled. Then you may have earned the spread rather than paid it, and every sign here flips. All of this models marketable orders.

The stock is cheap. At $2.50 a share, a 1-cent spread is 0.2% of price, not 0.01%. The percentages assume roughly $47. Recompute half-spread divided by your price.

What this doesn’t tell you

It measures one thing: the gap between your fill and a midpoint reference, times your share count.

It says nothing about whether the trade was worth making. A well-timed entry paying a 15-cent spread beats a badly-timed one paying a penny, and no execution analysis will ever capture that difference.

It does not extend to institutional size. Information leakage and the way large orders move the market against themselves are different problems with different mathematics.

It uses no broker’s published data. Every figure came from stated assumptions about spread and improvement rate. Your broker’s Rule 606 report shows which venues received your orders and what payments changed hands; execution quality data published under Rule 605 shows measured price improvement rather than my assumed rates. Those are your actual numbers, and they will not match these illustrations.

And it ignores tax entirely. Short-term versus long-term treatment on a gain usually swamps every figure in the turnover table.

FAQ

How much does payment for order flow cost me per trade?

Not directly measurable, because the payment flows from market maker to broker, not from you. What you can measure is your fill against the midpoint. On 200 shares filled at $47.033 with a $47.03 midpoint, that is $0.60. On 1,000 shares in a 10-cent-spread stock with no improvement at all, it is $50.00.

Is a zero-commission broker actually cheaper?

For small orders, yes, by arithmetic — a $6.95 flat fee on a $9,406 order was seven times the half-spread cost and it did not scale with size. For a 5,000-share order the commission was already only 0.003% of position value, so removing it changed little. You cross the spread under both models.

What percentage of my position does the spread cost?

Half the spread divided by the share price. A 2-cent spread on a $47 stock is 0.021%. The same 2-cent spread on a $5 stock is 0.20%, ten times more. A 25-cent spread on a $47 stock is 0.266%.

How often can I trade before spread costs matter?

On a $10,000 position with a 2-cent spread and 50% improvement, each round trip costs $2.12. Eleven round trips a year keeps you under 0.25% of position value. Weekly trading runs about 1.10% a year; roughly daily trading reaches 5.3%.

Does price improvement mean I got the best price available?

It means your fill beat the National Best Bid and Offer as of the instant the order was received. On 200 shares, 0.7 cents inside a 2-cent spread is $1.40 of improvement, and that figure is real. It is measured against one snapshot, so it does not rule out a better price on another venue or three hundred milliseconds later.

Why are option spreads so much worse in percentage terms?

Because the contract price is small relative to the tick. A contract at $1.20 by $1.30 has a 5-cent half-spread against a $1.25 midpoint, which is 4%. The same 10-cent spread on a $47 stock is 0.106%.

Do I pay the spread twice?

On a round trip, yes — once entering, once exiting. Every round-trip figure above already doubles the one-way cost, which is why a $1.06 crossing becomes $2.12.

Three things you can check today

Pull your broker’s Rule 606 report. It names the venues that received your orders and the payments received, broken out by order type.

Then find whatever execution quality data your broker publishes from its Rule 605 statistics. That gives you a measured improvement rate instead of the 50% I assumed.

Last, take your own order history and rerun the turnover table with your real number of round trips. That figure is the one that applies to you, and it is usually larger than people expect.

This article is general information, not financial advice. See our disclaimer.

Also worth reading: Dark Pools: Why Your 50,000-Share Order Hides

More in Market Mechanics · all topics · calculators · how this was checked