What Are the Best Benchmarks for Measuring RFQ Execution in Illiquid Markets?

Concept

The central challenge in measuring execution quality for illiquid assets is the absence of a continuous, reliable price reference. In liquid markets, a trader’s performance is judged against a dynamic tape of transaction data. For an off-the-run corporate bond, a niche derivative, or a block of restricted equity, this tape does not exist. The task, therefore, becomes one of creating a valid measurement event in a vacuum.

The Request for Quote (RFQ) protocol is the primary mechanism for achieving this. It is an engineered price discovery process, transforming a latent inquiry into a hard, competitive, and auditable data point.

The most effective benchmarks for this environment are those that measure the value created by the RFQ process itself. This requires a shift in perspective. Instead of comparing the execution price to a theoretical market-wide average that is likely stale or non-existent, the focus turns to measuring the quality of the competitive auction initiated by the request. The primary metric is price improvement relative to a defensible starting point.

This starting point, the arrival price, is the most critical pre-trade benchmark. It represents the state of the world at the moment the decision to transact is made, typically captured as the best available bid or offer (if one exists, however wide) or a recent indicative valuation.

A robust RFQ benchmark framework measures the quality of the price discovery event you create, not just the final execution level against a phantom market.

This approach reframes the problem from one of passive measurement to one of active performance evaluation. The quality of execution is a direct function of the structure of the RFQ. Factors such as the number of dealers invited, the timing of the request, and the information disclosed all contribute to the final outcome. Consequently, the benchmarks must capture these nuances.

They must quantify the spread between the winning and losing bids, the response times of the dealers, and the performance of the execution price against the arrival price. This creates a feedback loop where the data from each trade informs the strategy for the next, allowing for the systematic refinement of the execution process in markets defined by their opacity.

What Is the True Purpose of a Benchmark in Opaque Markets?

In opaque over-the-counter (OTC) markets, the purpose of a benchmark transcends simple performance scoring. Its true function is to construct a framework for decision-making under uncertainty. Given the absence of a central limit order book (CLOB), any single price point is suspect. A benchmark in this context serves as an anchor for a structured, repeatable process.

It provides the necessary tools to validate that the price achieved through a bilateral or multi-dealer negotiation was the best attainable price under the prevailing, albeit obscured, market conditions. This validation is essential for meeting regulatory obligations like MiFID II’s best execution requirements, which demand a demonstrable and auditable process.

Furthermore, these benchmarks provide a language for internal communication and strategy refinement. They allow a trading desk to quantify the value of its dealer relationships, to identify which counterparties provide the most competitive pricing in specific instruments, and to understand the trade-offs between information leakage and price competition. A well-designed benchmark system for illiquid RFQs is an intelligence-gathering operation. It systematically converts discrete trading events into a structured dataset that reveals the underlying liquidity network and informs future trading strategy with empirical evidence.

Strategy

A strategic approach to benchmarking RFQ execution in illiquid markets requires building a “benchmark waterfall.” This is a hierarchical framework of metrics, prioritized based on their relevance and the availability of data for a given asset. The strategy moves beyond a single-minded focus on the final price to a holistic evaluation of the entire execution process. The primary goal is to balance the competing pressures of achieving the best possible price improvement while minimizing the information leakage that could lead to adverse market impact.

At the top of this waterfall is the Arrival Price Benchmark. This is the foundational metric, representing the mid-price or a composite price at the moment the order is staged for execution. All subsequent performance is measured against this initial state. The second layer is Execution Spread Analysis , which measures the difference between the final execution price and the arrival price.

This is often expressed in basis points and is the most direct measure of the RFQ’s success. It answers the question ▴ “How much value did the competitive process generate?”

The third layer involves Dealer Performance Metrics. This moves from the individual trade to the strategic management of counterparty relationships. This analysis involves tracking metrics over time for each liquidity provider. Key indicators include:

• Win Rate ▴ The percentage of times a specific dealer provided the winning quote. A high win rate indicates consistent competitiveness.
• Response Time ▴ The average time it takes for a dealer to respond to an RFQ. Faster response times can be critical in time-sensitive situations.
• Quoted Spread ▴ The average bid-ask spread offered by a dealer. Tighter spreads generally indicate a greater willingness to take on risk.
• Price Improvement Contribution ▴ The amount of price improvement a specific dealer’s quotes contribute, even when they do not win the auction. This helps identify dealers who consistently provide competitive tension.

By systematically tracking these metrics, a trading desk can dynamically adjust its RFQ routing strategy, directing inquiries to the dealers most likely to provide the best liquidity and pricing for a specific instrument at a specific time. This data-driven approach transforms execution from a series of discrete events into a continuous process of strategic optimization.

Comparing Benchmark Methodologies

The choice of benchmark methodology is contingent on the specific characteristics of the asset and the market it trades in. A methodology that is effective for a slightly off-the-run sovereign bond will be inadequate for a highly structured credit derivative. The following table compares common benchmark methodologies and their applicability to illiquid RFQ execution.

How Should Information Leakage Be Factored into Strategy?

Information leakage is the implicit cost of signaling trading intent to the market. In the context of an RFQ, every dealer included in the request is a potential source of leakage. A dealer who receives the request but has no intention of winning the auction may still use the information to trade ahead of the requestor or to inform other market participants. This is a critical strategic consideration in illiquid markets, where even small trades can have a significant impact.

Minimizing information leakage is as vital as maximizing price improvement, as the former can easily negate any gains from the latter.

The primary strategy for managing this risk is the careful curation of the dealer list for each RFQ. Instead of broadcasting requests to a wide audience, a more surgical approach is required. This involves using the dealer performance metrics discussed earlier to identify a smaller, more competitive group of liquidity providers for each specific asset class.

A tiered system can be effective, where an initial RFQ is sent to a small group of trusted dealers, with the option to expand the request to a wider group if the initial responses are not competitive enough. This balances the need for competitive tension with the imperative to protect the confidentiality of the trade.

Execution

The execution of a robust benchmarking program for illiquid RFQs is a data-intensive process that transforms post-trade analysis into a pre-trade strategic asset. It requires a systematic approach to data capture, normalization, and analysis. The ultimate objective is to build a proprietary dataset that provides a clear, evidence-based view of execution quality and informs the continuous refinement of the firm’s trading protocols. This process moves beyond simply recording the winning price; it involves capturing the entire context of the RFQ event.

The first step is the implementation of a standardized data capture process for every RFQ. This can be integrated directly into the Order Management System (OMS) or Execution Management System (EMS). For each request, the system must log the following information:

1. Pre-Trade Data ▴ This includes the instrument identifier, the order size, the direction (buy/sell), the timestamp of the order’s creation, and the arrival price benchmark. The arrival price should be captured automatically from a pre-defined source, such as a composite pricing feed or an internal valuation model.
2. RFQ Protocol Data ▴ This covers the specifics of the RFQ itself, including the unique RFQ ID, the list of all dealers invited to quote, and the timestamp of the request’s dissemination.
3. Dealer Response Data ▴ For each dealer, the system must capture their response (or lack thereof), the bid and offer prices they provided, the size of their quote, and the timestamp of their response.
4. Execution Data ▴ This includes the dealer who won the auction, the final execution price and size, the execution timestamp, and any associated fees or commissions.

With this raw data captured, the next step is to process it through a Transaction Cost Analysis (TCA) engine. This engine calculates the key performance indicators (KPIs) that will be used to evaluate execution quality. The output of this analysis should be a detailed TCA report for each trade, as well as aggregated reports that provide a broader view of performance over time.

A Quantitative Model for RFQ Transaction Cost Analysis

A comprehensive TCA report for an illiquid RFQ must deconstruct the trade into its component costs and performance metrics. This allows the trading desk to pinpoint sources of alpha and friction in the execution process. The following table provides a template for such a report, using a hypothetical trade of an illiquid corporate bond as an example.

Fair Value Modeling in the Absence of Price

For the most illiquid securities, even a reliable arrival price may be unavailable. In these cases, a more advanced approach is required, centered on the concept of a “Fair Transfer Price” or “Micro-price”. This involves using a quantitative model to estimate the theoretical fair value of the asset at the time of the RFQ. This model can incorporate a variety of inputs, including:

• Comparable Instruments ▴ Prices of more liquid securities with similar characteristics (e.g. similar credit rating and duration for a bond).
• Factor Models ▴ The security’s sensitivity to broad market factors like interest rates, credit spreads, and volatility.
• Indicative Quotes ▴ Non-binding quotes from dealers or pricing services.

The output of this model provides a synthetic, defensible arrival price against which the RFQ execution can be measured. The goal is to execute at a price better than this modeled fair value. This approach introduces a higher level of analytical rigor and is particularly valuable for structured products and other complex instruments where traditional benchmarks fail entirely. It allows the firm to demonstrate a disciplined, model-driven approach to price discovery, even in the complete absence of observable market data.

Reflection

Calibrating the Execution Framework

The data and benchmarks derived from this analytical process are the raw materials for strategic refinement. They provide the necessary feedback to calibrate the firm’s execution framework. The insights gained from analyzing dealer performance, price improvement, and information leakage should directly influence the design of future RFQs. This creates a virtuous cycle ▴ each trade generates data, that data generates intelligence, and that intelligence sharpens the execution strategy for subsequent trades.

The ultimate goal is to build an operational architecture that is not static, but is a learning system, constantly adapting to the nuances of the market and the behavior of its participants. The question then becomes how this data-driven approach to RFQ execution can be integrated into the firm’s broader risk management and portfolio construction processes, transforming a specialized trading function into a source of firm-wide alpha.