University data sparks debate about trading policy

By Rachel Premack, Daily Staff Reporter
Published June 26, 2013

Economic data produced by the University in collaboration with Thomson Reuters has been in this month’s headlines of the Huffington Post, The Wall Street Journal and The New York Times — portraying the University in a less than positive light.

Traders pay an undisclosed fee to receive early access to a closely­ watched economic indicator compiled by the University, CNBC reported two weeks ago. Since then, everyone from members of the University community to New York Times op ­ed writer and economist Paul Krugman have raised questions concerning the morality and necessity of high­-frequency trading.

The figure, called the Index of Consumer Sentiment, measures consumer confidence and is part of the U.S. Department of Commerce’s Index of Leading Indicators.

University spokesman Rick Fitzgerald said Reuters has funded the research and handled its distribution since 2007.

Reuters provides “roughly” $1 million to the University for the research, Fitzgerald said. The agreement stipulates that Reuters distributes the data the University produces.

The monthly figure is published on Reuters’ website at 10 a.m., and Reuters’ paying customers receive the data early. For these traders, the data is distributed on a conference call at 9:55 a.m. or at 9:54:58.000 a.m. for subscribers to the ultra­-low latency distribution platform, according to Reuters.

Ultra-low latency is in a format specialized for high­-speed computerized trading, or high­-frequency trading. The CNBC report questioned the University's early release of data to paying traders, who were called members of a “select group ... on Wall Street” by AnnArbor.com.

Fitzgerald said the partnership with Reuters allows University researchers to focus on developing data. Previously, over 100 entities funded the research and the University distributed the index.

Michael Wellman, professor of electrical engineering and computer science who researches computational finance, said the University must always reconcile its responsibilities in the public and private sectors, the latter of which Fitzgerald said was needed to fund the Survey of Consumers.

“The University is engaging in this research that is ultimately provided to the public,” Wellman said. “It costs the University something to conduct the survey and create this indicator.”

He added that the early release of data is fair if parties are aware of the imbalance.

“The important thing is disclosure,” Wellman said. “If someone is getting a two second advantage, everyone should understand that.”

Fitzgerald said the options are fully disclosed in the two parties’ contract and are advertised to all customers. However, the CNBC report reflected that some users of the 9:55 a.m. conference call were unaware of the “ultra­-low latency” option two seconds prior. One called the undisclosed two-­second advantage “disingenuous.”

Uday Rajan, associate professor at the Ross School of Business, said even milliseconds are crucial in the “intense competition” of computerized trading.

“In each transaction, they make an extremely small amount of money,” Rajan said. “But the volume of transactions has become so huge in the last decade or so, it can add up to a large amount.”

Computerized trading has existed since the 1980s, Rajan said. High-­frequency trading, a 21st­ century invention, depends on sophisticated computers and algorithms that process information in milliseconds. Firms who use such technology can process large quantities of orders instantly.

A burst of trading activity was noted at 9:54:57.975 on May 17 by analysis firm Nanex. Within 10 milliseconds, more than 100,000 shares were exchanged and, 10 seconds in, $100 million were traded.

The two­-second advantage — and certainly the five­-minute one — may provide an advantage for traders whose stocks depend on consumer expectations for the market. If computers process that American consumers have low expectations for a salary raise, it may result in the selling of stocks. It is valuable for the trader that other market participants are unaware of consumer’s low expectations. In such a scenario, traders who purchased the 9:54:58.000 a.m. option view themselves as being in a better position than those who views consumer index information at 10 a.m.

Rajan said research has not concluded whether high-­frequency trading is beneficial or efficient, though he said the media stigmatizes it for the large quantity of money such traders may generate. He added that it provides a middleman for traders, thus lowering the fees needed for parties to trade shares.

However, Wellman said high­-frequency trading introduces unneeded costs, like specialized hardware that is necessary for traders to receive financial data at the exact millisecond.

Wellman authored a 2013 paper with Engineering doctoral candidate Elaine Wah that revealed the flaws of a certain type of high-­frequency trading called latency arbitrage.

Rajan said high-­frequency traders’ existence could lower participation with less market participants, meaning less competition.

“The other parties are aware that these guys can move very quickly and may be more reluctant to take part in trading,” he said. “That can be potentially harmful to other traders in the market.”