How can volatility in GBP/USD brokers be measured?

Direct answer

Volatility in GBP/USD “brokers” is best measured indirectly, by separating (1) market volatility in GBP/USD from (2) how a broker’s pricing, spreads, and execution create variability in the prices you effectively experience. In practice, you measure the variability of observed outcomes (quotes or fills) under a defined set of assumptions, then check whether the variability comes mostly from the market or from the brokerage layer.

Mechanism and definition: what volatility are you actually measuring?

“Volatility” usually means variability over time. Before measuring anything, define the object and unit of measurement:

Market volatility (base layer): variability of the GBP/USD exchange rate series over a time window (for example, minute-to-minute or day-to-day moves).
Broker-price variability (broker layer): variability in the bid/ask quotes you receive (or the mid-price you can trade) over the same time windows.
Execution variability (realized layer): variability in realized transaction prices (fills) relative to a reference, after including bid/ask spread effects and execution timing.

A common approach is to pick a reference price and compute returns (percentage or log changes) from a time series. Then compute volatility as the dispersion of returns. For example, if you have a sequence of returns (r_t) over (N) intervals, one standard measure is the sample standard deviation of (r_t). If you compute this from the market series, you measure market volatility; if you compute it from broker-provided quote or fill series, you measure broker-influenced variability.

Evidence or example: measurement choices you can independently replicate

Here is a concrete, non-price-predictive way to compare variability using only saved data you collect or already have access to.

1) Choose a time horizon and sampling rule

Pick an interval length (e.g., 1 minute) and use a consistent sampling method:

Quote sampling: record broker bid and ask at each interval boundary.
Fill sampling: record fills and compute realized returns per trade (requires trade timestamps and a reference).

Assumption: you are analyzing the same horizons across series.

2) Use more than one dispersion metric

Standard deviation of returns is common, but it may hide distribution shape. Using multiple metrics improves interpretability:

Standard deviation of returns (volatility magnitude).
Median absolute deviation (MAD) of returns (robust to outliers).
Interquartile range (IQR) of returns (another dispersion view).

Assumption: the metrics are computed on aligned timestamps and the same transformation (e.g., log returns).

3) Separate “market move” from “cost and execution” using a spread-aware decomposition

Even without any live pricing claims, you can structure a decomposition conceptually:

Mid-price return: based on the quote mid (mid = (bid+ask)/2).
Effective price gap: related to how far the execution price is from the mid or reference.

If you observe that broker-level volatility spikes occur mainly when bid/ask spread widens, that points to broker-layer variability driven by costs/liquidity conditions rather than changes in underlying market volatility.

Possible computation pattern (conceptual):

Measure volatility of mid returns (market-like component).
Measure variability of effective gap (cost/execution component).
Compare their contributions by analyzing which component’s dispersion rises most when overall variability rises.

4) Run a simple sanity check on stationarity

Volatility estimates can break when market regimes change. Use two windows (early vs late) and compare whether your chosen volatility measure remains similar. Large shifts suggest regime changes or data-generation changes.

Limitations and risks: why broker volatility measurement can fail

At least one material limitation matters for any broker-related volatility measurement:

Changing microstructure conditions: spreads and liquidity can vary throughout the day, so volatility in broker-observed prices may reflect changing trading conditions rather than a stable “broker volatility.”
Data definition risk: quote series and fill series are not interchangeable. Fills depend on order type, latency, and availability of counterparties; quotes may not reflect what is achievable at decision time.
Reference mismatch: if your reference (e.g., mid price, external rate, or last traded price) is not aligned in time and definition, your computed “returns” and effective gaps become biased.
Non-synchronous sampling: if you sample quotes at fixed times but fills occur irregularly, you may introduce artificial variability.
Historical relationships don’t transfer: a volatility estimate from one period does not establish what will happen later.