"Market Microstructure 101: How Prices Are Really Formed"

Classical finance treats price as given: assets have values, values move with information, and markets simply report the result. Market microstructure — the field that studies how trading actually happens — starts from a more honest premise: price is not reported, it is produced, by a specific mechanism, under specific frictions, one order at a time. Understanding that production process is the difference between trading a market and trading a chart of it.

Price discovery is a search process

At any moment there is a hypothetical "efficient price" — the value that would prevail if all information were aggregated instantly and trading were free. No one observes it. What the market displays instead is a bid, an ask, and a last trade, each contaminated by frictions: the spread, the discreteness of ticks, the temporary pressure of large orders, the inventories of intermediaries.

Microstructure's central insight is that markets search for the efficient price through order flow. Every trade is a small experiment: someone was willing to cross the spread, and everyone else updates. Hasbrouck (2007) formalized this by decomposing price changes into a permanent component (information) and a transitory component (friction and noise). Practically, this means the last traded price is a noisy estimate — and much of short-horizon "price action" is the noise, not the signal.

The three characters of every market

Microstructure theory populates markets with three archetypes, and the interaction among them generates prices:

Informed traders possess information not yet in the price — a superior forecast, faster processing of public news, or genuine private knowledge. They trade because they expect the price to move toward their estimate.

Uninformed (liquidity) traders trade for reasons unrelated to short-term value: hedging, rebalancing, cash needs. They pay the spread as a fee for immediacy.

Market makers stand between the two, quoting both sides continuously. Their problem is elegant and brutal: they cannot tell which incoming order is informed. Every quote they post is an option they've written to someone who may know more than they do.

Kyle's (1985) model showed how an informed trader optimally hides inside uninformed flow, trading gradually so the market maker cannot distinguish signal from noise — and how, in the process, information leaks into price trade by trade. This is price discovery in its rawest form: not an announcement, but a slow surrender of an informational edge through the act of trading on it.

Order flow is the carrier of information

The operational consequence: signed order flow — the net of buyer-initiated versus seller-initiated volume — is the single most informative variable in short-horizon price formation. Decades of empirical work confirm that order-flow imbalance predicts price changes at short horizons far better than past prices do, because flow is the mechanism by which private information becomes public price.

This reframes several things traders take for granted:

• "Support" and "resistance" are, mechanically, concentrations of resting limit orders. They hold when the book absorbs incoming flow and break when it doesn't.
• Momentum at short horizons is often the visible half of a large order being worked through the book — impact first, information later, if at all.
• Volume without imbalance moves nothing. Two-sided volume is intermediation; one-sided volume is information (or forced flow), and the book prices them differently.

Frictions are not imperfections — they are the market

Tick sizes, spreads, latency, queue priority, and fees are usually described as imperfections layered on top of a "true" market. Microstructure inverts this: the frictions are the market design, and they shape behavior as strongly as information does. A coarser tick size deepens queues and raises the value of time priority. A maker-taker fee schedule reroutes flow. A change in matching algorithm — FIFO versus pro-rata — reorganizes the entire ecology of participants in a product.

For systematic traders this carries a hard lesson: a strategy is never just a forecast. It is a forecast plus an interaction with a specific market design, and the second term often dominates the first at short horizons. Backtests that model prices but not frictions systematically overstate what the forecast is worth — a failure mode we dissect in "Why Most Backtests Lie."

What this means in practice

Three durable takeaways for anyone operating in modern electronic markets, futures very much included:

1. Trade the mechanism, not the abstraction. Know your product's tick size, matching rules, typical depth, and session liquidity profile. These parameters set the physics your strategy lives in.
2. Expect to be modeled. Your order flow is someone else's signal. Execution that leaks intent — predictable sizing, rhythmic timing — pays a tax to whoever reads it.
3. Respect the transitory component. Much of what looks like opportunity at short horizons is friction that reverts. Infrastructure that timestamps and records every order event lets you measure, after the fact, how much of your edge was information and how much was noise you paid for — which is precisely why disciplined execution logging is an analytical asset, not merely a compliance obligation.

Prices are made, not found. The traders who internalize that stop asking what the market is saying and start asking what the mechanism is doing.

References

• Kyle, A. S. (1985). "Continuous Auctions and Insider Trading." Econometrica, 53(6).
• O'Hara, M. (1995). Market Microstructure Theory. Blackwell.
• Hasbrouck, J. (2007). Empirical Market Microstructure. Oxford University Press.
• Harris, L. (2003). Trading and Exchanges. Oxford University Press.