FX spot rate blending methods
Table of Contents
Summary
When several source feeds are available, spot rate blending can combine them in several ways: a simple average, a weighted average (weights reflecting feed quality, latency, or liquidity relationships), a median or trimmed mean (robust to a stale or erroneous outlier), or a waterfall / governed selection (a preferred source with a fallback hierarchy). The choice is a governance decision owned by Finance / Market Risk, and changing it requires sign-off.
Detail
The methods
- Simple average — arithmetic mean of all available source rates. Easiest; no opinion about relative source quality.
- Weighted average — sources carry weights, often reflecting feed quality, latency, or the bank's own liquidity relationships; higher-quality or more timely feeds get higher weights.
- Median / trimmed mean — robust statistics that suppress outliers from a stale or erroneous feed without manual intervention.
- Waterfall / governed selection — a preferred source is used when present and in tolerance; a defined fallback hierarchy is invoked otherwise.
Governance, not code
The methodology is set by Finance or Market Risk and is defensible and auditable; method changes require approval. The same multi-source statistics (count, average, median, min/max, standard deviation) used for emerging-market broker-quote IPV on rates and vols can in principle be applied to spot, though spot's observability (Level 1) means it rarely needs the heavier treatment.
Industry precedent
Robust, multi-source, statistically-defensible rate construction is the direction of travel for FX benchmarks generally — the WM/Reuters fix moved to a volume-weighted median over a widened fixing window precisely to resist manipulation and outliers (see References).
See also
- Spot rate blending — the parent concept.
- Spot rate governance — checks the blended output.
References
- FSB Foreign Exchange Benchmarks — Final Report (Financial Stability Board, Sep 2014) — WM/Reuters reform: wider fixing window, volume-weighted median. https://www.fsb.org/publications/r_140930.pdf
- WM/Refinitiv FX Benchmarks — Overview (LSEG/FTSE Russell) — current volume-weighted-median methodology and treatment of illiquid pairs. https://www.lseg.com/content/dam/ftse-russell/en_us/documents/factsheets/wmr-overview-and-request-for-feedback.pdf
- To fix or not to fix: representativeness of the WM/R methodology (Breedon & Ranaldo, 2024, J. Int. Financial Markets) — argues for further robustness in multi-source weighting. https://www.sciencedirect.com/science/article/pii/S0927538X24000623
- Foreign Exchange Market Microstructure and the WM/Reuters 4pm Fix (Ito & Yamada, 2015) — why fix-window methodology matters for derived spot.
- Fixing the Fix? (FCA Occasional Paper No. 46, 2018) — empirical assessment of post-reform fix methodology. https://www.fca.org.uk/publication/occasional-papers/occasional-paper-46.pdf