Transaction Cost Analysis (TCA)

The direct, observable charges itemized on a trade confirmation. These are the hard costs of accessing the market infrastructure.

• Commissions: Fees paid to the executing broker for their service, typically on a per-share or basis-point basis.
• Exchange Fees: Charges levied by the trading venue for matching orders, often structured under a maker-taker model where rebates are paid for adding liquidity and fees are charged for removing it.
• Clearing & Settlement Fees: Costs imposed by central counterparties (CCPs) and custodians for guaranteeing and finalizing the transfer of ownership.
• Regulatory Taxes: Government-imposed transaction taxes, such as the UK Stamp Duty or the French Financial Transaction Tax.

While the most visible, explicit costs are often dwarfed by implicit costs for large institutional orders.

The non-observable friction arising from the interaction of the order with the market. These costs are inferred, not invoiced, and represent the true challenge of execution.

• Market Impact: The adverse price movement caused by the trade's own supply/demand pressure. It has a permanent component (information leakage) and a temporary component (liquidity concession).
• Spread Cost: The cost of crossing the bid-ask spread, measured by the effective spread (2 * |Trade Price - Midpoint|).
• Delay Cost: The slippage between the arrival price (when the decision was made) and the price when the order is first sent, reflecting the risk of waiting.
• Opportunity Cost: The forgone profit from an unfilled order. If a buy order misses a rally, the cost is the paper gain that wasn't realized.

The forecasting of expected transaction costs before an order is released, used to calibrate strategy parameters and set realistic expectations.

• Cost Curves: Mathematical models that map expected market impact as a non-linear function of order size relative to average daily volume (ADV), volatility, and urgency. A common model is the square-root impact rule.
• Liquidity Profiling: Analyzing real-time volume profiles and order book depth to identify high-liquidity nodes where large orders can be absorbed with minimal impact.
• Strategy Selection: Using the pre-trade cost estimate to choose between a passive POV (Percent of Volume) algo, a schedule-driven TWAP, or an aggressive liquidity-seeking algo.
• Venue Analysis: Projecting fill probabilities and adverse selection risk across lit exchanges and dark pools to optimize the Smart Order Router (SOR) configuration.

The raw data feeding TCA is contaminated by market microstructure effects that must be filtered to avoid drawing false conclusions.

• Bid-Ask Bounce: Transaction prices oscillating between bid and ask create spurious negative autocorrelation in returns. Using midpoint prices instead of trade prices mitigates this distortion.
• Tick Size Constraints: The minimum price increment (tick size) discretizes the price grid, creating rounding effects that bias spread cost estimates, especially for low-priced securities.
• Timestamp Synchronization: Latency between the trade report and the quote used for benchmarking can create phantom costs. Precision timestamping (microsecond granularity) is critical for high-frequency TCA.
• Outlier Filtering: Erroneous prints, block trades executed under special rules, and off-market transfers must be algorithmically cleansed before analysis to prevent skewing the cost distribution.

The gold standard benchmark for measuring total transaction cost. It captures the difference between the decision price (paper portfolio return) and the actual execution price (real portfolio return).

• Formula: Implementation Shortfall = (Paper Return) - (Actual Return)
• Decomposes into: Delay Cost + Market Impact + Opportunity Cost + Explicit Commissions
• Captures the cost of not trading as well as the cost of trading

The adverse price movement caused by your own order consuming resting liquidity. It represents the premium paid for immediacy and the information leakage of your trading intent.

• Temporary Impact: Transient price pressure that dissipates as liquidity replenishes
• Permanent Impact: Lasting price change reflecting the information content of the trade
• Modeled using the Square Root Law: Impact ≈ σ · √(Q/V) where Q is order size and V is average daily volume

Two foundational execution benchmarks that slice time differently:

• VWAP (Volume Weighted Average Price): Weights each price by the volume traded at that price. Formula: Σ(Price × Volume) / Σ(Volume). Ideal for comparing execution against the market's natural rhythm.
• TWAP (Time Weighted Average Price): Averages prices at fixed time intervals, ignoring volume. Formula: Σ(Price) / N intervals. Used when volume distribution is unpredictable.
• VWAP is the dominant benchmark for institutional TCA; TWAP is preferred for illiquid securities

Measures the round-trip cost of a transaction by comparing the trade price to the prevailing midpoint at the time of execution.

• Formula: 2 × |Trade Price - Midpoint|
• Captures the implicit cost of crossing the bid-ask spread
• A trade at the bid implies a buyer-initiated transaction; at the ask implies seller-initiated
• Used to detect payment for order flow quality and adverse selection

The cost of trading against a counterparty with superior information. After your trade, the price moves permanently against you because the counterparty knew something you didn't.

• Measured by the permanent price change 5-10 minutes post-trade
• High adverse selection signals toxic order flow
• Probability of Informed Trading (PIN) models estimate this risk pre-trade
• Market makers widen spreads to compensate for this risk

A systematic broker algorithm allocation mechanism that randomly routes parent orders across a pre-approved set of execution algorithms.

• Purpose: Eliminates trader bias and enables statistically valid A/B testing
• Post-trade TCA scores feed back into the wheel to dynamically re-weight allocations
• Underperforming algos are down-weighted; top performers receive more flow
• Creates a continuous optimization loop for execution quality