Direct answer
To assess “slippage around news” in a defensible way, you need inputs that (1) define the news event timing, (2) measure trade execution outcomes, (3) establish a reference price for comparison, and (4) capture costs and data quality. The goal is not to predict outcomes, but to calculate observed differences and understand where uncertainty comes from.
Mechanism and definition
Slippage usually means the gap between a chosen reference price at (or near) order submission and the actual execution price (or the effective average price for a fill). When the market reacts to news, volatility and liquidity can change rapidly, so the reference-to-execution gap can widen.
A complete assessment typically distinguishes stable mechanics from variable conditions:
- Market movement: price changes caused by new information.
- Execution friction: liquidity constraints, bid–ask spread widening, partial fills, and the effect of order routing/handling.
- Costs: explicit costs (if any) and implicit costs reflected in spreads and fees.
Data inputs to collect (and why)
1) News event specification
Collect a clear event identity and timestamp:
- Event time: when the data release is scheduled or when you define “the start” of the reaction.
- Time standard and timezone: the same standard across all data sources.
- Event window: a rule for what “around news” means (for example, from X seconds before to Y seconds after). State the rule even if you later test multiple windows.
2) Reference price construction
Choose and document how you define the reference price used for slippage:
- Reference source: mid-price, last traded price, or bid/ask midpoint from a specified venue.
- Reference timestamp rule: whether it is at order submission, at the nearest tick, or sampled at a fixed frequency.
- Consistency rule: use the same method for every observation.
3) Execution and order data
You need enough detail to reconstruct what happened:
- Order submission time and order parameters (side, order type, size).
- Fill timestamps and fill prices (or at least an effective average fill price).
- Partial fill indicators: whether the position was built over multiple fills.
- Canceled/expired orders: to avoid mixing “attempts” and “successful executions” without tracking.
4) Cost and liquidity proxies
To separate friction from pure price movement, include variables such as:
- Bid–ask spread and its time series near the event window.
- Liquidity/availability proxies available in your data (for example, depth measures if you have them).
- Any fees or commission components you can observe, applied using the same accounting rule.
5) Data provenance and transformation log
Create a short audit trail:
- Where each field comes from (platform export, market data feed, calendar source).
- Any time conversions, resampling, or filtering rules.
- How you handle missing values (drop rows, interpolate, or label as missing).
Evidence or example calculation (with explicit assumptions)
A typical calculation defines slippage per fill as:
- Slippage = (effective execution price) − (reference price), with sign defined consistently for buys vs sells.
One workable approach is to aggregate slippage over an event window:
- Compute slippage using the reference price at order submission time (assumption).
- For each event, average slippage across all fills that fall within your reaction window (assumption).
- Also compute the distribution (median and percentiles) because news-driven effects can be heavy-tailed.
If you test multiple window sizes, treat window choice as a variable and report results for each window rather than selecting one after seeing the outcome.
Limitations and risks (material failure modes)
At least one material limitation matters for most studies:
- Timestamp mismatch: event times, market-data timestamps, and execution timestamps may be in different timezones or have different latency. Even small alignment errors can look like slippage.
- Reference price bias: choosing a reference that lags or leads the true decision moment can overstate or understate slippage.
- Partial fills and order behavior: if fills occur over many seconds during volatility spikes, “slippage” may mix market impact with execution timing.
- Selection bias: filtering to only filled orders can remove the worst outcomes and make results look cleaner.
- Non-stationarity: relationships observed historically do not guarantee similar behavior in future news or different regimes.
Verification and next questions
To verify your assessment, apply quality checks before interpreting effects:
- Confirm timestamp alignment by checking whether pre-event spreads and price movements look plausible. - Quantify data loss (missing fills, missing quotes, or incomplete orders).