Direct answer: why liquidity and spreads can be misleading
Liquidity and spreads are often discussed as if they were stable, observable inputs. In practice, both are conditional. Liquidity can thin out or concentrate depending on time, volatility, and order-flow, and spreads can widen even when quoted prices look similar. That means a single “liquidity level” or “spread size” can stop being a reliable representation of what will happen during execution.
Mechanism and definition: what the concepts actually mean
Liquidity is the ease with which an asset can be bought or sold near a given price, without moving that price too much. It is not one number; it is typically reflected by things like available depth, trading volume, and how quickly orders appear and disappear.
Spreads are the difference between the best available buy (bid) and sell (ask) prices at a moment in time. A quoted spread is a snapshot. Real transaction costs are not only the quoted spread; they also include execution effects such as slippage (getting filled at worse prices than expected) and any additional fees or operational costs.
To discuss limitations clearly, separate stable mechanics from variable conditions:
- Stable mechanic: “spread” and “liquidity” are defined through matching of bids and asks.
- Variable conditions: liquidity depth, volatility, and participant behavior can change rapidly.
Evidence or example: how assumptions break
Consider a simple cost approximation: expected execution cost increases when the spread widens. This only works if a few assumptions are true: the spread remains near its snapshot value, the order can be filled without materially worsening, and trading venue and execution pathway remain consistent.
If any assumption changes, the approximation fails. For example, if volatility rises, order books can become less stable, widening spreads and increasing the probability that fills occur after price moves. If liquidity is concentrated in small sizes, an order that is larger than the visible depth may “walk the book,” moving through multiple price levels. If the data used to compute liquidity is not synchronized with the execution moment, you are comparing different realities (a common limitation when using delayed or averaged observations).
Limitations and risks: failure modes to watch for
-
Liquidity is time-dependent Liquidity is not permanently present at the same “quality.” It can be abundant in one period and scarce in another, especially around news releases, abrupt regime changes, or when fewer participants provide quotes.
-
Quoted spreads can differ from realized costs A tight quoted spread does not guarantee low realized cost. If your order size is large relative to available depth or if execution is slower than expected, you can still experience higher total cost through slippage.
-
Relationships can be non-stationary Even if you observe that “higher liquidity historically correlates with narrower spreads,” that relationship may not persist during new market conditions. Historical relationships do not ensure future results, particularly when volatility dynamics or participant composition changes.
-
Provider and venue differences “Liquidity” and “spread” can be measured differently across venues and platforms. Different matching rules, quoting behavior, and aggregation methods can produce different observed spreads even for the same underlying market concept.
-
Model sensitivity and definitions Any calculation depends on definitions (what counts as liquidity depth, which time window is used, how often quotes are sampled). Small differences in definitions can produce large differences in conclusions.
Verification or next question: how to check reliability
To independently verify claims about liquidity and spreads, focus on what can be measured and on transparent assumptions:
- Use the same time basis for observation and evaluation (avoid comparing delayed snapshots with execution-time expectations).
- Distinguish quoted spread from realized execution cost by separating the two in your analysis.
- Specify order size assumptions when relating liquidity to expected execution.
- Check whether observed liquidity–spread patterns persist across different market regimes rather than relying on a single historical window.
A useful next question is not “what is the spread” or “how liquid is it,” but: under what conditions would the observed measures stop matching execution reality?