L1 liquidity

The bid-ask spread is universally understood as a transaction cost. Less appreciated: it's the market maker's real-time estimate of adverse selection risk, and its dynamics encode the informational state of the market.

The Glosten-Milgrom decomposition

Glosten & Milgrom (1985) proved that the bid-ask spread emerges endogenously in a market with asymmetric information. Even a competitive, zero-profit market maker must charge a spread because some fraction of incoming orders are from informed traders who know the asset's true value. The spread compensates for the expected losses to informed counterparties.

Huang & Stoll (1997) decomposed the spread empirically into three components:

1. Adverse selection cost — compensation for trading against informed participants
2. Inventory holding cost — compensation for carrying unwanted inventory risk
3. Order processing cost — the operational cost of market making

Their estimation showed that adverse selection accounts for a substantial fraction (often 30–50%) of observed spreads in equity markets. In crypto markets, where informed participants can be algorithmic and fast, the adverse selection component is likely even larger.

What we measure — and what it reveals

• Mean spread (bps) — the average transaction cost environment. But averages hide regime changes.
• Spread standard deviation — this is the underappreciated metric. A bar with mean spread of 2 bps could have been consistently at 2 bps (stable market making, confident fair-value estimates) or oscillated between 0.5 and 8 bps (episodic uncertainty, potential information events). Bollerslev & Melvin (1994) showed that spread volatility itself predicts future return volatility.
• Spread extremes — the minimum and maximum spread within the bar. Maximum spread captures the moment when market makers were most defensive — often coinciding with large informed trades or exogenous news.
• Spread-to-depth ratio — spread normalized by available depth. A tight spread on thin depth is cheap but fragile; a tight spread on substantial depth is genuine liquidity.

The information in spread dynamics

Spread widening is a leading indicator. When market makers widen spreads, they're signaling increased uncertainty about fair value — either because they've observed toxic flow (Easley, López de Prado & O'Hara, 2012), or because volatility is elevated and their hedging lags reality.

Conversely, spread compression signals a market maker competition regime: multiple liquidity providers are confident in their fair-value estimates and competing for priority by narrowing quotes. This is the environment where mean-reversion strategies are most reliable and execution costs are lowest.

The L1 Liquidity dataset pairs spread metrics with depth at the best quote — because the two dimensions together define the true execution landscape. A tight spread with thin depth invites market impact; a tight spread with substantial depth is the holy grail of execution environments.

References:

• Glosten, L.R. & Milgrom, P.R. (1985). "Bid, Ask and Transaction Prices in a Specialist Market with Heterogeneously Informed Traders." Journal of Financial Economics, 14(1), 71–100.
• Huang, R.D. & Stoll, H.R. (1997). "The Components of the Bid-Ask Spread: A General Approach." Review of Financial Studies, 10(4), 995–1034.
• Bollerslev, T. & Melvin, M. (1994). "Bid-Ask Spreads and Volatility in the Foreign Exchange Market." Journal of International Economics, 36(3-4), 355–372.
• Easley, D., López de Prado, M.M. & O'Hara, M. (2012). "Flow Toxicity and Liquidity in a High-Frequency World." Review of Financial Studies, 25(5), 1457–1493.