How Algorithmic Trading Affects Market Liquidity

Algorithmic trading has grown to represent a substantial share of overall trading volume across both traditional and digital currency markets, and its effects on market liquidity are more nuanced than either purely positive or purely negative characterizations suggest.

Under normal market conditions, algorithmic trading, particularly algorithmic market-making strategies discussed in earlier articles, generally contributes positively to market liquidity. Automated market-making systems can continuously provide buy and sell quotes with tighter spreads than would typically be achievable through manual trading, since algorithms can process and react to changing conditions far faster than human traders, allowing for more precise and responsive pricing that benefits all market participants through reduced transaction costs.

Algorithmic trading has also generally increased overall trading volume and market participation, since automated strategies can identify and act on smaller, more fleeting opportunities that wouldn't be practical for a human trader to consistently capture, adding to overall market activity and, generally, contributing to deeper, more liquid markets during typical trading conditions.

However, research and documented market events have also identified circumstances where algorithmic trading can contribute to reduced liquidity, particularly during periods of market stress. Many algorithmic market-making strategies include risk management logic designed to reduce activity or withdraw entirely during periods of extreme volatility or unusual market conditions, precisely to protect the algorithm's own risk parameters. When many market participants employ similar risk management logic simultaneously, this can result in a rapid, collective withdrawal of liquidity precisely when it's most needed, exacerbating price volatility during already stressed conditions rather than providing the stabilizing function that liquidity providers typically serve during calmer periods.

This dynamic has been studied extensively in the context of various flash crash events across both traditional and digital currency markets, where sudden, severe price movements were amplified, at least in part, by the simultaneous withdrawal of algorithmic liquidity providers reacting to rapidly changing conditions, sometimes followed by cascading effects as other automated systems reacted to the resulting price movements in turn.

For investors, understanding this nuanced relationship between algorithmic trading and liquidity has practical implications. Markets may appear highly liquid and efficient under normal conditions, but that liquidity isn't necessarily a permanent, structural feature that can be relied upon during periods of genuine stress. This is particularly relevant for large trades or positions in less established digital assets, where the liquidity available during calm market conditions may not accurately represent the liquidity that would actually be available during a broader market downturn, when many algorithmic liquidity providers might simultaneously reduce their activity.

This underscores the practical value of not assuming current, observed liquidity conditions will necessarily persist during periods of market stress, and building appropriate margins of safety into position sizing and risk management accordingly, particularly for less liquid assets where this dynamic tends to be most pronounced.