In this paper, we empirically investigate the short-term impact of human/algorithmic limit order submissions on the liquidity provision and withdrawal process of other human/algorithmic traders. Using a high-frequency dataset containing over 1.5 million limit orders in the USD/JPY and EUR/JPY foreign exchange spot markets (amounting to a limit order volume of approximately $2 trillion), we document three key findings. First, order-splitting strategies widely adopted by algorithmic traders to disguise the true order size seem to go detected and are perceived as more information-rich or predatory than orders of the corresponding size typically submitted by human traders. Second, the inverse relationship between limit order size and price aggressiveness is less consistent than expected – both concerning traders’ strategic order submissions and their impact on the liquidity withdrawal by others. Third, we find that traders appear to be more sensitive to limit orders submitted from the same side (non-execution risk) than to the opposite side of the order book (free option risk), but that the ‘recovery’ of the limit order book primarily is driven by a reassessment of free option risk.
|Publication status||Published - Jan 2018|