Market Microstructure and Price Formation

Market Microstructure: How Prices Are Formed
Market microstructure studies the mechanisms of price formation, the role of information, the behavior of participants, and the influence of trading rules on market quality. This is a relatively young area of financial science, which began developing in the 1970s and became especially relevant with the advent of electronic trading.

Information and Price Formation
Prices reflect information available to market participants. The efficient market hypothesis asserts that prices instantly incorporate all relevant information. Microstructure examines exactly how this process takes place. Informed traders possess private information about the fundamental value of the asset. Their trading gradually reveals information to the market — prices move toward the true value. The speed of revelation depends on the intensity of trading by informed agents. Uninformed (liquidity) traders trade for reasons unrelated to information — portfolio rebalancing, cash needs. They create noise, complicating the extraction of the information signal from prices.

Kyle Model
The Kyle model (1985) is a fundamental model of informed trading. The informed trader knows the true value of the asset. The market-maker observes the aggregate order flow but cannot distinguish informed orders from uninformed ones. Equilibrium: the informed trader trades gradually, so as not to reveal information too quickly. The market-maker sets prices linearly dependent on order flow — the parameter $\lambda$ (lambda) measures the impact of orders on price. The key conclusion: the spread and price impact reflect the informational content of trading. In markets with a greater share of informed trading, spreads are wider.

Glosten-Milgrom Model
The Glosten-Milgrom model analyzes the spread from the viewpoint of adverse selection. The market-maker quotes bid and ask, not knowing whom they are trading with. If the counterparty is informed, the market-maker incurs losses — buys before a price decline or sells before a rise. The spread compensates for expected losses from trading with informed participants. The higher the probability of informed trading, the wider the spread. This explains why spreads widen before important news — the probability of the presence of informed agents increases.

High-Frequency Trading (HFT)
High-frequency trading uses speed as a competitive advantage. HFT firms place servers near exchange servers (colocation), use specialized equipment and algorithms to execute trades in microseconds. The role of HFT in microstructure is ambiguous. On one hand, HFT improves liquidity — spreads have narrowed, market depth has increased. On the other hand, HFT can amplify volatility in stressful situations, when algorithms synchronously withdraw from the market. The Flash Crash on May 6, 2010 demonstrated the risks: the Dow Jones index fell by 9% within minutes, then recovered. Investigation showed that algorithm interactions created a cascade of sales.

Market Fragmentation
Modern markets are fragmented among multiple venues. In the US, stocks trade on NYSE, NASDAQ, multiple ECNs, and dark pools. Regulation (Reg NMS) requires execution at the best available price (NBBO), but allows trading on any venue. Fragmentation creates opportunities for arbitrage between venues and complicates liquidity aggregation. Smart Order Routing (SOR) algorithms direct orders to venues with the best conditions.

Practical Consequences
Understanding microstructure helps optimize execution. Large orders should be split and executed gradually (TWAP, VWAP, Implementation Shortfall algorithms). Choice of order type affects costs — limit orders save on the spread but carry the risk of non-execution. Microstructure analysis can provide informational signals. Order imbalance, spread changes, activity in dark pools — all this reflects the actions of informed participants.