1. Ask two different questions
Volatility concerns the variability of a specified series over a specified horizon. Market liquidity concerns the conditions under which transactions can be completed. These dimensions may interact, but one is not a definition of the other. A quiet series does not establish that a substantial transaction could be completed easily, and active movement does not automatically imply an absence of trading capacity. A useful state description therefore identifies the measurements behind both axes. This guide explains research concepts and hypothetical classifications, not investment advice or assessments of any current market.
2. Specify the volatility estimator
For a realized-volatility measure, record the return definition, sampling interval, window length, and treatment of missing observations. State whether overnight and intraday movements are both included. If dispersion is annualized, disclose the scaling convention and its assumptions instead of treating the multiplier as a law. An option-implied measure answers a different question and should not silently replace a realized measure. A classification threshold calibrated for one horizon or estimator has no automatic meaning for another. Retain the raw measure alongside the categorical label so the transformation remains inspectable.
3. Treat market liquidity as multidimensional
BIS work on liquidity measurement distinguishes trading cost, capacity, and resilience. A spread-based measure captures one aspect of transaction conditions; depth addresses available quantity; resilience concerns recovery after disruption. None alone describes every order size, venue, or execution schedule. Choose proxies that match the research question and disclose their limitations. Indicative observations are not promises of executable terms. A broad state label should not imply that every constituent instrument has identical liquidity, especially when a composite blends markets with different trading conventions and observation coverage.
4. Separate funding capacity from tradability
Funding liquidity concerns the ability to meet financing and payment needs; market liquidity concerns transacting in assets. BIS research on market conditions treats funding and market liquidity as distinct dimensions alongside volatility. A researcher studying funding should identify the financing variable being observed rather than substitute a trading-volume series. Conversely, an asset's active turnover does not establish a participant's ability to fund a position. Keep these concepts in separate fields even when the eventual framework combines them, and avoid claiming a causal channel merely because the measures move together.
5. Check whether observations can conceal stress
A low measured movement rate can reflect genuinely small changes, but a researcher should also test for stale observations, sparse trading, and aggregation choices. High recorded activity can coexist with poor conditions for a particular transaction size. Use timestamp freshness, sample coverage, and session matching to challenge apparently reassuring measurements. Do not replace missing liquidity data with a favorable state. A classification can remain mathematically reproducible while economically misleading if its inputs describe different sessions, different market segments, or different degrees of observability.
6. Read a fictional two-axis grid
The invented grid uses two dimensionless scores solely to demonstrate classification. Higher V denotes greater measured variability; higher L denotes greater measured liquidity impairment. Both thresholds are 1.0, with equality assigned to the higher category. Case A is quiet but impaired, while case B is variable without crossing the impairment threshold. These combinations show why collapsing the axes into a single reassuring or alarming word loses information. The scores have no empirical calibration, probability interpretation, or execution meaning, and their thresholds are not proposed market standards.
7. Preserve scope in the final label
A useful record names the market segment, measurement interval, estimator, liquidity proxy, threshold version, and input status. Include whether the observation was contemporaneous, corrected later, or reconstructed after the event. If combining indicators, disclose weights and how missing components affect the result. Validate the measurement process before interpreting changes as shifts in underlying conditions. The label can organize a review of market functioning, but it cannot establish future volatility, guaranteed transaction capacity, or a preferred position. Those are separate claims requiring their own definitions and evidence.
Inspect the comparison
| Fictional case | V score | L impairment score | Threshold classification |
|---|---|---|---|
| A | 0.4 | 1.3 | Lower variability; higher impairment |
| B | 1.4 | 0.6 | Higher variability; lower impairment |
| C | 1.0 | 1.0 | Higher variability; higher impairment |
| D | 0.7 | Missing | Liquidity state unavailable |
Primary references
References support the stated methods and data conventions. Original fictional examples are not observations from the cited sources.