How the Major U.S. Stock Markets Actually Work
When you tap "buy" on a brokerage app, your order does not go to a single building called "the stock market." It gets routed through a competitive web of exchanges, market makers, and dark venues, and where it lands can quietly affect the price you pay. Understanding that structure explains both why U.S. markets are the deepest in the world and why execution quality is not automatic.
The structure behind "the market"
The New York Stock Exchange and the Nasdaq are the two names most investors know, but they are only the visible tip of a much larger system. U.S. equities trade across roughly sixteen registered stock exchanges, several dozen alternative trading systems (mostly dark pools), and hundreds of internal market-making desks operated by broker-dealers. A single share of a large company like a major bank or a widely held index fund can legally trade on any of these venues at any moment, and the exchanges compete openly for that order flow through pricing, speed, and rebates.
The NYSE still uses a hybrid model built around a designated market maker, a firm obligated to maintain a fair and orderly market in specific stocks by posting bids and offers and, when necessary, trading against the trend to cushion imbalances. Nasdaq, by contrast, has always been a pure electronic dealer market: multiple competing market makers post quotes for the same stock, and a matching engine executes against the best price available. Both models have converged over the past two decades toward high speed, computer driven execution, but the underlying philosophy, a single accountable specialist versus many competing dealers, still shapes how each handles a large or unusual order.
Layered on top of the exchanges is a category of trading that never appears on a public order book at all: the over-the-counter and dark pool segment. Roughly 40 to 45 percent of U.S. equity trading volume in a typical year executes off the lit exchanges entirely, most of it through wholesale market makers that internalize retail orders (fill them directly from their own inventory rather than sending them to an exchange) and through institutional dark pools that let large investors trade blocks of stock without revealing their intentions ahead of time. This matters because the price you see quoted on an exchange is not necessarily the price a huge share of actual trades take place at.
The math: NBBO and where your fill actually comes from
The rule that stitches the fragmented system together is called the National Best Bid and Offer, or NBBO, established under Regulation NMS in 2005. Every exchange must publish its best bid and ask in real time, a central system aggregates them, and no venue is allowed to execute a trade at a price worse than the best bid or ask available anywhere in the network at that instant. This is the single most important piece of market plumbing most investors have never heard of.
Worked example 1: how the NBBO is built. Suppose a stock is quoted simultaneously on three venues:
- Exchange A: bid $74.10, ask $74.16 (spread $0.06)
- Exchange B: bid $74.12, ask $74.18 (spread $0.06)
- Exchange C: bid $74.09, ask $74.14 (spread $0.05)
The NBBO takes the highest bid across all three, $74.12 from Exchange B, and the lowest ask across all three, $74.14 from Exchange C. The consolidated NBBO spread is therefore $74.14 minus $74.12 equals $0.02, narrower than any single exchange's own quoted spread of $0.05 or $0.06. This is the direct, measurable payoff of competition among venues: no single exchange has to be the best on both sides for the investor to get the best available combined price. A market order to buy must be filled at $74.14 or better, regardless of which of the sixteen exchanges receives it.
Worked example 2: what payment for order flow actually costs, and saves, you. Most commission-free brokers route retail market orders not to an exchange but to a wholesale market maker, which pays the broker a small fee for the right to fill the order internally, typically a fraction of a cent per share. Suppose the NBBO on a stock is bid $50.02, ask $50.06, so the midpoint is $50.04. A wholesaler receives your order to buy 300 shares and fills it at $50.05, one cent better than the public ask, while paying your broker $0.0015 per share for the order.
Your price improvement: ($50.06 minus $50.05) times 300 shares equals $3.00 saved versus the publicly quoted ask. The broker's payment for order flow revenue on your trade: $0.0015 times 300 equals $0.45. Both numbers are real, and they are not in conflict: the wholesaler can profit from the spread between what it pays retail traders in aggregate and what it earns trading against institutional flow, while still giving you a better price than the public quote and paying your broker a commission substitute. The academic and regulatory debate over payment for order flow is not about whether individual retail trades get price improvement, data consistently shows most do, it is about whether a different routing arrangement, without the payment, would produce even better prices still. Reasonable analysts disagree on the size of that gap; it is measured in fractions of a penny per share on a typical order, not dollars.
What the evidence shows about market quality
Market microstructure research since Regulation NMS took effect has converged on a few consistent findings. First, quoted and effective spreads on large, liquid U.S. stocks have fallen dramatically since the shift to electronic and fragmented trading in the 2000s, from several cents per share in the specialist era to often a penny or less on the most actively traded names today. Second, this improvement has not been uniform: small-cap and thinly traded stocks still carry meaningfully wider spreads and higher price impact, because market makers demand more compensation for holding inventory in names with less trading activity and less predictable order flow.
Third, the rise of high frequency, low latency trading has been a genuinely two sided development in the empirical literature. Studies using order book data generally find that faster, more competitive market making has narrowed spreads and deepened quoted liquidity in normal conditions, benefiting nearly all investors who trade during regular hours. The same research also documents that liquidity can evaporate very quickly during stress events, most visibly during the flash crash of May 2010, when major indexes fell and then recovered a large share of the loss within minutes as automated liquidity providers withdrew simultaneously. The lesson from two decades of data is not that fragmentation and speed are bad for investors, the aggregate evidence says they are not, but that market structure resilience during stress is a distinct problem from market structure efficiency during calm conditions, and both remain ongoing subjects of regulatory attention.
It is worth being precise about who benefits most from this fragmentation, because the answer differs by investor type. Large institutional traders, pension funds and mutual funds moving blocks worth tens of millions of dollars, rely on dark pools specifically to avoid signaling their intentions; a visible order to buy a million shares would itself move the price against the buyer before the order finished filling, a cost known as market impact. Retail investors trading a few hundred shares rarely move prices at all, so the dark pool ecosystem exists mostly to solve a problem retail investors do not have. What retail investors do inherit from the same fragmented structure is the NBBO protection and the wholesale market maker competition described below, both of which tend to work in their favor precisely because their order sizes are small and predictable.
What this means for your orders
For a long-term investor buying a diversified index fund a few times a month, none of this changes the outcome in any way that matters. NBBO protection means you are, by rule, getting at least the best publicly available price at the moment of execution, and the spread on a widely held fund is usually a fraction of a cent relative to share price. The market structure story becomes practically relevant in three situations: trading in thinly traded small-cap or micro-cap stocks, where spreads and price impact are large enough to matter; placing sizable orders relative to a stock's typical daily volume, where market impact costs rise nonlinearly; and trading during volatile or fast moving markets, where displayed liquidity can thin out faster than a market order can react.
In each of those situations, the tool that neutralizes most of the structural risk is the same: a limit order, which specifies the worst price you are willing to accept, rather than a market order, which accepts whatever price is available the instant it arrives. A limit order cannot protect you from a stock simply not trading at your price, but it guarantees you will never be filled at a price you did not choose, which is precisely the protection you want during a volatile open, a low liquidity stock, or a fast market.
There is a second, quieter implication for anyone building a portfolio around index funds rather than individual stocks: the deep, competitive structure described above is exactly why broad market index funds can be traded and rebalanced so cheaply at scale. A fund tracking a broad U.S. index trades in the most liquid names in the entire system, where NBBO spreads are often a fraction of a penny and price impact from routine rebalancing flows is negligible. That structural cheapness is one of the quieter reasons index investing has become steadily less expensive over the past two decades, on top of the well known decline in expense ratios themselves.
Actionable breakdown
- Use limit orders for any stock with wide spreads or thin volume.
- Trust the NBBO rule; you cannot legally be filled worse.
- Avoid market orders in the first and last minutes of trading.
- Do not fear payment for order flow on liquid, large-cap names.
- Size orders relative to average daily volume for illiquid stocks.
- Split very large orders over time instead of one market order.
Common pitfalls
The most common mistake is placing a market order in a thinly traded stock at the open or close, when displayed liquidity is thinnest and prices can gap sharply within seconds; a limit order costs nothing and removes this risk entirely. A second pitfall is assuming that because your broker is commission free, execution quality does not matter; the difference between routing arrangements is usually small per share but compounds across a lifetime of trades. A third is confusing exchange volume with total market activity, since a large share of trading never touches a public exchange at all, which can distort an investor's sense of how liquid a given stock really is. A fourth is treating flash crash style events as evidence that markets are broken, when the more accurate reading of the data is that normal condition liquidity and stress condition liquidity are genuinely different phenomena.
The bottom line
U.S. equity markets are a competitive, fragmented, and heavily regulated network rather than a single exchange, and that structure has made routine trading cheaper for nearly everyone while leaving small-cap liquidity and stress period resilience as the two places where it still pays to be careful.
All articles · How markets work · New trading strategies · Trading costs · Regulation of securities markets