Executive Summary:
Stock markets around the world-Euronext, Toronto, Madrid, and Australia, to name a few- reduce transparency by providing traders with the option to hide some or all of their bets, their "limit orders." These "iceberg" or "reserve" orders, where some of the size is revealed and the rest a mystery, presents a problem for buy-side institutions and trading desks at bankruptcy-pending Lehman Brothers, Goldman Sachs, and others. The paper "Hidden Liquidity: An Analysis of Order Exposure Strategies in Electronic Stock Markets" by SMU Cox finance professor Kumar Venkataraman and co-authors Bessembinder and Panayides offers insights for policymakers, buy- and sell-side institutions, and designers of stock markets.
Background
Recently presented at the Financial Industry Regulatory Authority (FINRA) and discussed at the Securities and Exchange Commission (SEC) with economists, this new research reveals the first evidence on the costs and benefits of hiding order size in electronic markets. Hidden orders, while useful and necessary, come with headaches for institutions and brokers trying to find the other side: they need to predict the hidden quantity of shares behind a displayed order of say, 1000. Are there 30,000 more shares hidden behind it, or 5,000-or is the order for 1,000 simply 1,000? The problem is more acute in the U.S. after Regulation National Market System was phased in; it is expected to become more acute in Europe since the implementation of MiFID.
Author Kumar Venkataraman says, "With Reg NMS, two trends are significantly complicating trading problems. One is the further fragmentation of the centralized markets for trading NYSE and Nasdaq stocks. The second is the explosion in trading platforms with reserve orders or dark liquidity pools. If all liquidity is displayed, re-aggregating the liquidity across fragmented markets is a relatively trivial problem. But with hidden quantity at each price point, or with completely dark liquidity pools, it becomes a daunting task; one needs to predict not just whether reserve size exists but also how much at each price point. In other words, you need to understand when traders hide size and how much, and then reverse engineer the hiding decision."
The study presents an approach to detect the reserve quantity based on the characteristics of stock being traded, the underlying market conditions, and observable trading intentions (whether informed or uninformed). One striking finding is that hidden liquidity can be discerned by how aggressive or passive the order is being priced.
"The job of an institutional trading desk is to collect as much information as possible and identify signals useful for reverse engineering," Venkataraman explains. "Next, they need to determine how much weight to give each signal, and how to aggregate information to come up with a prediction." He continues, "This is exactly what we do: we show when traders hide size and for what types of stocks. And for each order, we come up with a Hidden-Score, the probability the order is hidden, and a Hidden-Size, our best guess of the reserve quantity, which trading desks can use to design optimal order routing strategies.
While the techniques used by broker-dealers and algorithmic trading platforms are proprietary, as academics, the authors study the same problem from the perspective of trader behavior and seek to publish their findings. "We show that hidden liquidity can be predicted to a significant extent based on many market observables," Venkataraman states.
Reserve (iceberg) orders are a common but under-studied market feature, according to Venkataraman. One version of the hidden order is the crossing networks-the so-called "dark liquidity pools" that don't require users to display price quotes. Dark pools have been very successful in drawing away large traders from traditional exchanges. "In our research, we examine iceberg or reserve orders where some portion needs to be exposed," Venkataraman clarifies. "In dark pools, orders are routed to a black box, metaphorically speaking, fingers-crossed. Our research provides some clues on when you might have a better chance of finding liquidity in dark pools." Dark pools will soon come under increasing regulatory scrutiny, predicts Venkataraman. "We know little about how dark pools affect price discovery or other market quality measures, which is a concern."
To Hide or Not to Hide
The study relies on Euronext's order-level stock trading data from 2003, with information on the displayed and hidden quantity for every order in every stock. The authors find that hidden orders are used extensively on Euronext, accounting for 18% of incoming orders and 44% of the order volume. Orders tend to be hidden more often in less liquid stocks, about 50% of the time, and in larger order sizes, about 70% for orders greater than 50,000 euros. In their findings, hidden orders were shown to be less likely to execute completely, and exhibited longer times to execute. The price risk of a delayed or unexecuted trade is the cost to the trader for hiding order size. They document that the benefit to the trader for hiding order size is lower trading costs.
So who uses hidden orders? Based on the findings, the authors conclude that informed traders, who are concerned about the risk of delayed execution, tend to place aggressively priced orders that are fully displayed so as to draw out counterparties (or possibly reactive traders) and obtain faster execution. In contrast, larger, less aggressively priced orders tend to be hidden. Reserve orders are used primarily by uninformed traders to mitigate the option value of a standing limit order (e.g. buy only at $40) and lower price impact of front-running by parasitic traders.
Other tendencies to hide or not hide orders were observed in the study. Hidden orders are more likely when spreads are wide, suggesting the tendency to hide orders when information asymmetry is high. Orders are less likely to be hidden for liquid stocks, less volatile stocks, and stocks with larger relative tick size. Though larger total order size is associated with a greater likelihood to hide, traders who displayed larger size are less likely to hide a portion of their order. Fully displayed orders were however more likely to execute completely, a tangible benefit of transparency. Traders can draw reactive traders by either posting an attractive price or by exposing order size. The two methods of attracting reactive traders differ in their relative costs and benefits. A more aggressive order gains price priority over orders at inferior prices, while a fully exposed order gains time priority versus hidden orders at the same price.
Does reverse engineering work? The presence and magnitude of reserve orders can be predicted to a significant extent based on the attributes of orders, firm characteristics and market conditions. However, the ability to detect hidden size is less than perfect suggesting that reserve orders allow large traders to partially conceal their trading intentions. Thus, the study concludes that hidden orders are an important risk control tool for market participants. When considering whether to expose order size, traders face both costs and benefits of doing so. Exposing an order increases the chance that it will attract counterparties. On the other hand, exposing an order could cause other traders to withdraw liquidity, or employ front-running strategies.
Implications
That hidden orders are used extensively and strategically on Euronext Paris implies that these orders are valuable to traders. Thus, market centers with reserve orders may be more successful in attracting large orders. The NYSE, in particular, currently allows only floor brokers to use reserve orders. Findings are suggestive that the steep drop in NYSE's market share and the trend toward institution-oriented "dark pools" can be attributed at least in part to design weaknesses at the NYSE. The set of order types that traders can submit represents an important dimension of trading system design.
In an increasingly fragmented and "dark" U.S. market place, institutional trading desks, responsible for executing block orders received from portfolio managers, are facing new challenges in the search for liquidity pools. The study provides an approach for trading desks to detect reserve size and also identifies circumstances when liquidity is likely to be more hidden.
The study has implications for regulators. On one hand, the market fragmentation and reduced book transparency has the potential to lower market quality and price efficiency in financial marketplace. Alternatively, the study suggests that reducing transparency by allowing traders to hide a portion (or all) of order size may help exchanges consolidate order flow in a centralized marketplace.
In a world pushing for transparency in all things financial, there's a time and a place for the partial shade provided in hidden liquidity.
"Hidden Liquidity: An Analysis of Order Exposure Strategies in Electronic Stock Markets" by Kumar Venkataraman of Southern Methodist University's Cox School of Business with Hendrik Bessembinder and Marios Panayides of University of Utah is under review, and was presented/discussed at the FINRA and the SEC in July 2008.
Written by Jennifer Warren. |
