Who likes large uninformed orders? Anyone?
Increased intermediation has led to large institutional investors who occasionally demand liquidity beyond what a market can offer instantly. These investors resort to splitting their order to trade small (child) orders optimally through time, often constrained by a time limit. How does such algorithmic execution of an uninformed whale order affect other market participations, such as market makers and other investors?
Before sharing with you our insights developed by laboring through an economic model, you might be interested to pause a moment, think, and share your intuition by answering the single question stated below. If you do so, you will be able to compare your intuition with that of others (as you will see how others responded in a pie chart, all anonymous).
The model we set up to answer this question features strategic (i.e., endogenous) liquidity supply and demand. In a Stackelberg game a large institutional seller has to trade before a particular time limit d (for duration). He will do so optimally, fully cognizant of the response of (high-frequency) market makers. The latter supply liquidity optimally by setting bid and ask prices to earn off of the spread, yet mean-revert out of costly non-zero inventory. Other investors arrive randomly and trade optimally at the quotes they observe when arriving at the market. (Importantly, all market participants are equally informed on fundamental value and the friction therefore is a pure allocational one.)
After working through the math and calibrating the model to trading in blue-chip stocks here is our short answer: All benefit! The gains from trade are enjoyed by the large seller, the market makers, and other investors. There is a catch though. The analysis reveals that this result is not generally true. The more subtle message is:
- Market makers benefit only if risk capacity of the market is large enough and duration of the order is sufficiently short (exact condition are in the paper).
- Other investors benefit only if the large seller trades at high enough intensity. These investors effectively enjoy the low prices (caused by price pressures) that market makers set to mean-revert out of inventory. Yet, they pay the elevated bid-ask spread at the time when the large seller is selling. The former effect dominates the latter when the large seller's trade intensity is high enough (again, precise results are in the paper).
Finally, the analysis reveals that knowledge of how long the large seller's order will last, benefits all. The intuition is that revelation of d makes optimal inventory control easier for market makers and since the latter are competitive the benefits are (partially) passed on to the large seller. The intuition is that, all else equal, market makers willingly accommodate more of the large sell order if they can predict when it ends. The model yields precise (analytic) results on such sunshine trading as compared to stealth trading (where d remains hidden and is therefore private to the large seller).
Importantly, the model is tractable and yields analytic results. This should help discussing market quality and regulating markets that increasingly feature whale orders.
P.S.: The paper is with Agostino Capponi and Hongzhong Zhang of Columbia University (New York). Please find it here.