How Can a Tca Framework Differentiate between Slippage Caused by Market Impact and Last Look Rejections?

Concept

An execution policy’s success is measured by its ability to translate a portfolio manager’s alpha into realized returns. The interface between the order and the market is where this translation occurs, and slippage is the primary metric of its efficiency. From a systems perspective, slippage represents the friction within the market’s architecture ▴ a composite of costs incurred to transform a trading decision into a completed position.

Disaggregating this composite cost is the foundational task of a modern Transaction Cost Analysis (TCA) framework. The system must isolate two fundamentally different sources of friction ▴ the intrinsic cost of liquidity consumption, known as market impact, and the extrinsic cost imposed by a counterparty’s discretionary power, most acutely observed in last look rejections.

Market impact is an inherent property of the market structure itself. It is the price concession a participant must make to incentivize others to take the opposite side of a trade, particularly a large one, within a specific timeframe. This is a physical constraint of the system, akin to the water displaced by a ship. The size and speed of the ship dictate the size of the wake.

Similarly, the order’s size relative to available liquidity and the urgency of its execution determine the magnitude of the market impact. It is a predictable, modelable cost of doing business in a world of finite liquidity. A sophisticated TCA framework approaches market impact as a problem of physics and measurement, quantifying the price degradation directly attributable to the order’s own footprint.

A TCA framework’s primary conceptual function is to deconstruct the total cost of execution into its constituent parts, assigning causality to either market physics or counterparty behavior.

Last look rejections introduce a different class of problem. This phenomenon is a feature of certain market designs, particularly within foreign exchange and some over-the-counter environments. It grants a liquidity provider (LP) a final, optional moment to renege on a quoted price after a client has committed to trading. A rejection is the LP exercising this option.

The resulting slippage arises from what happens next ▴ the client must go back to the market, which may have moved adversely in the intervening milliseconds, and attempt to execute again. This is a cost born of game theory, not market physics. The LP is making a calculated decision based on its own risk position and the short-term trajectory of the market. The resulting cost to the client is a direct transfer of risk and a consequence of the LP’s informational advantage during the “last look” window. Differentiating this from market impact requires the TCA system to move beyond simple price measurement and into the realm of event analysis, focusing on counterparty behavior, timing, and the context of the rejection itself.

What Is the Core Architectural Difference in Measurement?

The architectural divergence in measuring these two costs is profound. To measure market impact, the TCA system requires a high-fidelity snapshot of the order book and prevailing market prices at the precise moment of order arrival (T0). The analysis then compares the execution prices of the subsequent child orders against this T0 benchmark.

The deviation, adjusted for general market movements, provides a measure of impact. It is an analysis of price changes relative to a baseline state.

To measure the cost of last look, the system must function as a state machine, tracking the entire lifecycle of a request. It logs the time the quote was requested, the price that was quoted, the time the trade was submitted, the time the rejection message was received, and the market price at the moment of rejection. The cost is calculated by comparing the original quoted price to the price at which the trade was eventually filled elsewhere. This analysis is less about a single price benchmark and more about a sequence of events and the time elapsed between them.

It is an investigation of a broken trade, a forensic analysis of a specific counterparty’s decision and its financial consequences. The system must capture not just prices, but also the metadata of the interaction ▴ the rejection code, the hold time, and the identity of the rejecting counterparty.

Strategy

A strategic TCA framework designed to unbundle market impact from last look costs operates on a principle of targeted data segmentation and contextual analysis. The goal is to create two distinct analytical pathways, each with its own set of metrics and benchmarks, that allow for the clear attribution of slippage. The overarching strategy is to first classify every component of an execution ▴ every child order, whether filled or rejected ▴ by its interaction type and venue, and then apply a specialized lens to each category.

The first strategic pillar is Liquidity Source Classification. The TCA system must ingest and parse execution data with a full understanding of the venue on which the trade was attempted. This involves maintaining a detailed map of the execution ecosystem, categorizing each liquidity pool as either “firm” or “last look.” Firm liquidity venues, like a traditional exchange’s central limit order book, provide certainty of execution for any order that crosses the spread. On these venues, last look is not a variable, so any slippage beyond the bid-ask spread can be primarily attributed to market impact or timing discrepancies.

Last look venues, common in FX, operate on a quote-driven model where LPs have discretion. By separating all trading activity into these two buckets from the outset, the analysis can immediately isolate the trades where last look is a potential cause of cost.

Effective strategic differentiation requires classifying execution venues by their liquidity model ▴ firm or last look ▴ before applying specific analytical metrics to each.

Developing a Dichotomous Metrics Framework

With data segmented by liquidity type, the second strategic pillar is the application of a Dichotomous Metrics Framework. This involves using one set of metrics tailored to diagnose market impact and a separate, distinct set designed to expose the costs of last look. The two are not interchangeable. Comparing them side-by-side for a given set of parent orders reveals the dominant sources of friction.

For analyzing market impact, typically on firm venues, the strategy focuses on price-centric metrics benchmarked against the state of the market at the time of the parent order’s arrival. The key is to measure the price degradation caused by the order’s own demand for liquidity.

• Arrival Price Slippage ▴ This is the foundational metric. It measures the difference between the average execution price and the market midpoint at the moment the parent order was received by the execution algorithm. A high slippage value for a large order on a firm venue is a strong indicator of market impact.
• Reversion Analysis (Markouts) ▴ After a fill, the TCA system analyzes the market price over subsequent short intervals (e.g. 1 second, 5 seconds, 30 seconds). If the price tends to revert after a buy order (i.e. move back down), it suggests the buy order temporarily pushed the price up, a classic sign of market impact. A permanent price move, conversely, suggests the order was trading in the direction of new information.
• Execution Shortfall Implementation ▴ This framework compares the final portfolio value against the value of a hypothetical paper portfolio where all trades were executed at the arrival price. This provides a holistic measure of all execution costs, with market impact being a primary component.

For analyzing last look costs, the strategy shifts from price degradation to event-driven and time-based metrics. The goal is to quantify the cost of the optionality given to the liquidity provider.

• Rejection Rate Analysis ▴ The most basic metric is the percentage of child orders sent to a last look venue that are rejected. This is often analyzed on a per-LP basis. A high rejection rate is the first sign that last look is a significant factor in execution quality.
• Hold Time Measurement ▴ This is the time elapsed between when a trade is sent to an LP and when a response (fill or reject) is received. Sophisticated TCA systems measure this in milliseconds. Long hold times, especially on rejected orders, are a major red flag. During this hold time, the LP can observe market movements and decide whether to fill the order, a practice known as “free optionality.”
• Post-Rejection Cost Analysis ▴ This is the most direct way to measure the financial harm of a rejection. The system calculates the difference between the price of the rejected quote and the price at which the order was eventually filled elsewhere. This “cost of delay” can be substantial if the market moves unfavorably during the time it takes to reroute the order.

How Do These Strategies Combine in Practice?

The power of this dichotomous approach is realized when the two sets of analytics are viewed in concert. A portfolio manager might see a high overall slippage figure for a large currency order. The TCA framework first breaks this down by venue. It might show that the portion of the order executed on firm ECNs had moderate slippage consistent with market impact models.

In parallel, it would show the portion routed to last look banks. Here, the analysis might reveal a 40% rejection rate from one particular LP, with an average hold time of 250 milliseconds on rejects. The post-rejection cost analysis then shows that after each rejection, the eventual fill price was, on average, 0.5 pips worse. This allows for a precise diagnosis ▴ the total slippage was not just a function of the order’s size, but was significantly inflated by the behavior of a specific counterparty, a cost attributable to the last look mechanism.

The following table illustrates how the TCA framework presents this strategic differentiation:

Execution

The execution of a TCA strategy capable of dissecting slippage requires a robust technological and data-analytical architecture. It is a process of capturing high-resolution data, applying precise and standardized calculations, and presenting the results in a diagnostically useful manner. This is where the theoretical strategy meets the operational reality of market data feeds, protocol messages, and database queries. The entire system is predicated on the ability to reconstruct the full lifecycle of every parent and child order with microsecond precision.

The Data Capture and Analysis Pipeline

The operational flow begins with the capture of every relevant message connected to an order’s journey. This is a non-trivial data engineering challenge.

1. Order and Market Data Logging ▴ The core of the system is an event logger that captures and timestamps every FIX (Financial Information eXchange) message sent and received by the firm’s Order Management System (OMS) or Execution Management System (EMS). Simultaneously, a separate process must be capturing and timestamping the direct market data feed from the relevant venues. Crucially, these two sets of timestamps must be synchronized to the same clock, often using Network Time Protocol (NTP), to allow for meaningful comparison.
2. Data Normalization and Enrichment ▴ Raw FIX messages and market data are cryptic. A normalization layer parses this data into a structured database format. This process enriches the data by, for example, attaching the liquidity type (firm/last look) to the venue identifier, or calculating the prevailing bid-ask spread at the time of each event from the market data stream.
3. Event Reconstruction ▴ The system then reconstructs the “story” of each child order. It links the NewOrderSingle message to its corresponding ExecutionReport messages (which could be acknowledgements, partial fills, full fills, or rejections). This creates a complete, time-ordered sequence of events for every single attempt to trade.
4. Benchmark Calculation and Slippage Attribution ▴ With the event history reconstructed, the analytical engine applies the dichotomous metrics. For a child order that was filled, it calculates arrival price slippage. For a child order that was rejected, it calculates the hold time and flags it for post-rejection cost analysis. The system then aggregates these child-level analytics up to the parent order level, providing a consolidated view.

Leveraging the FIX Protocol for Deeper Insights

The FIX protocol is the language of electronic trading, and within its tag-value pair structure lies the specific data needed to differentiate slippage sources. A properly configured TCA system must capture and interpret these specific tags.

A granular analysis of FIX protocol messages is the definitive method for capturing the event-level data required to distinguish between market dynamics and counterparty actions.

By capturing these tags for every child order, the TCA system can build a forensic log. When an ExecutionReport with OrdStatus=8 (Rejected) is received for an order sent to a known last look venue, the system immediately flags it. It calculates the hold time by subtracting the time the order was sent from the TransactTime on the rejection message. It then monitors the ClOrdID to see where that portion of the order is next routed and filled, allowing it to calculate the post-rejection cost.

A Practical Example of Execution Analysis

Consider a parent order to buy 100 million EUR/USD. The execution algorithm splits this into ten child orders of 10 million each. The TCA system logs the following (simplified) sequence for one of those child orders:

• 10:00:00.100 ▴ Child Order #5 (Buy 10M EUR/USD) is sent to LP ‘BANK-A’ (a known last look provider) at a quoted price of 1.0850.
• 10:00:00.450 ▴ The system receives an ExecutionReport for Child Order #5 from BANK-A. OrdStatus=8 (Rejected). TransactTime=10:00:00.445.
• 10:00:00.460 ▴ The execution algorithm immediately reroutes a new Child Order #5b (Buy 10M EUR/USD) to a firm ECN.
• 10:00:00.475 ▴ Child Order #5b is filled on the ECN at an average price of 1.0851.

The TCA execution analysis performs the following calculations:

1. Hold Time ▴ 10:00:00.445 (TransactTime) – 10:00:00.100 (Time Sent) = 345 milliseconds. This is a significant hold time.
2. Rejection Event ▴ The order is flagged as a last look rejection from BANK-A.
3. Post-Rejection Cost ▴ 1.0851 (Actual Fill Price) – 1.0850 (Rejected Price) = 0.0001, or 1 pip. For a 10 million EUR order, this single rejection cost the firm $1,000.

This analysis is then aggregated with the results for all other child orders. If several orders sent to BANK-A show a similar pattern, the system can definitively state that a significant portion of the total slippage for the parent order was caused by last look rejections from that specific counterparty. Simultaneously, for child orders routed to firm venues, the system would perform a standard market impact analysis, comparing their fill prices to the arrival price benchmark at 10:00:00.000. This provides the complete, bifurcated picture of execution costs.

Reflection

The ability to architect a system that distinguishes between the physics of market impact and the gamesmanship of last look provides more than just a clearer P&L. It transforms the execution process from a passive experience into an active, data-driven strategy. The framework detailed here is a diagnostic engine, and its output is a map of the hidden costs and opportunities within your execution network. The data reveals which counterparties are true partners and which are imposing undue friction. It quantifies the precise cost of a discretionary pause, turning a vague frustration into a hard number that can inform routing decisions and counterparty negotiations.

Ultimately, this level of analytical granularity serves a higher purpose. It is about control. By understanding the true source of every basis point of slippage, you gain the ability to intelligently manage it.

This knowledge informs the design of smarter routing logic, provides the leverage needed for productive conversations with liquidity providers, and ultimately, creates a more resilient and efficient pathway for translating investment ideas into realized alpha. The question then becomes, how is your current operational framework architected to capture and act upon this critical intelligence?