Study highlights US trading venue order handling disparities
The differences in how order types are handled across displayed US market centres can significantly impact execution performance, according to a recent analysis from consultancy Woodbine Associates.
The study, ”US exchange and ECN performance, second-half 2009', examines the price improvement and mean reversion, i.e. the performance of a stock after an order has been executed, of market orders and marketable limit orders, as well as the realised spread associated with displayed limit orders across US execution venues.
It concludes that executions on BATS Exchange, Direct Edge and NYSE Arca Equities provided the best “balanced” environment in terms of price improvement for market and marketable limit orders, without facilitating adverse selection for liquidity providers.
Matt Samelson, principal at Woodbine and author of the report, suggests that technology may be a significant contributor to execution performance on the venues in question.
“From the results, it seems as though a high standard of technology combined with significant volume can contribute to a more balanced market for traders, with negligible information advantage for either liquidity takers or providers,” Samelson told theTRADEnews.com. “The report shows that market participants should not consider visible costs such as take fees and rebates in isolation, but also the degree of price improvement and information imbalances on each venue.”
The study also found that the New York Stock Exchange (NYSE) provided the best price improvement for marketable limit orders, i.e. those that stipulate a maximum price for execution based on the last matched trade, but left participants that display liquidity at risk of adverse selection.
According to Samelson, this supports NYSE's recent move from operating a hybrid floor and electronic trading platform model to appointing designated market makers, i.e. firms on the floor that have an obligation to provide liquidity in NYSE-listed stocks. The switch to the DMM model occurred roughly two-thirds of the way into the data set.
“Most people had concerns that the hybrid model allowed specialists to get an advanced look at the order book and in theory manipulate the market,” said Samelson. “The data seems to support this. Liquidity providers at the New York Stock Exchange, on average, appear to have been subjected to adverse selection, suggesting that specialists may have traded against this liquidity at times most favorable to them but least favorable to the liquidity provider.”
Woodbine Associates analysed data that US regulator the Securities and Exchange Commission requires exchanges, ECNs and alternative trading systems to send to them on a monthly basis. It relates to the use of five order types – market orders, marketable limit orders, inside-the-quote limit orders, at-the-quote limit orders and near-the-quote limit orders – across all US exchanges and alternative trading venues.
Samelson notes that some venues have questioned analyses based on this data set, claiming that the standardisation process compromises its ability to capture variations in execution performance between different venues. Furthermore, the data does not include opening and closing auction data and orders that have special handling instructions. But he adds that if used correctly the data can provide value, particularly for price improvement and mean reversion statistics, and a more relevant analysis of US market centres than has currently been the focus.
“Up until now, there has been a large focus on dark pool liquidity, and the benefits associates with non-displayed execution, yet this accounts for only about 8% of average daily market volume on a single-counted basis,” he said. “Given that basic market and marketable limit orders reflect approximately 48% of displayed market prints and displayed limit orders account for some 65% of prints, overlooking this flow could be a substantial oversight in evaluating the quality of a firm's trading. While some have suggested that the difference between displayed venues is negligible, it is measurable and does make a difference.”