In the wake of the May 6 stock market “flash” crash, in which the U.S. stock market lost $862 billion in value in a matter of minutes — only to recover approximately $600 billion of that loss within half an hour — regulators have been scrambling to figure out what may have caused the crash and how it could be prevented in the future.Two Cornell professors may have found an answer.Prof. Maureen O’Hara, finance, and her husband, Prof. David Easley, economics, who chairs the department, have developed a statistical model that may help predict the potential for future stock market crashes. Their model, which they named Volume Synchronized Informed Trading, predicted the flash crash two hours before it happened.“VPIN moved 2 hours before that crash. It moves in advance of price movements; VPIN has some predictive power,” O’Hara said.The model monitors market makers’ activity within the market. It measures the shares of “informed traders” –– which include sophisticated computer models and market makers –– shares in the market place versus those held by uninformed traders, according to O’Hara.When there are too many informed traders within the marketplace relative to uninformed ones, market makers will curtail their activity in the market in order to avoid losing money to what they perceive as better informed sources. This, in turn, has the potential to lead to a crash, as the market makers provide the market with needed liquidity. “VPIN is a method to measure imbalances in trade … it’s a warning signal of potential illiquidity” said O’Hara.The model relies on public data detailing the gap between buying and selling stock shares and compares it to the market’s overall trade volume. A wide disparity could forewarn the presence of high levels of informed trading, or toxic flow from sources that are more informed about short-term price movements.If, however, market makers had this information in advance, they would be able to make more calculated trading decisions, rather than curtailing all of their trading activity suddenly, which could lead to market crashes.“Markets are high-frequency, so the risks are also fast. What you need are measurements that can warn you ahead of time as to what will happen — it can reduce liquidity risks by being predictive and help our markets,” Easley said.“VPIN measure is a way to estimate what fraction or order is toxic … it’s insurance against toxic flow” O’Hara added.On Oct. 15, O’Hara and Easley filed a patent with Tudor Investment Corporation for their formula.The two professors have worked at Cornell for over 30 years. Experts in their fields, O’Hara’s area of concentration includes “banking and security markets,” while Easley concentrates on “economic theories and pricing equilibrium models.”O’Hara said that they are currently focusing on macro-level market structures.“We look in detail how prices are set and evolve — a look at what markets actually do,” she said.The professors hoped their work would resonate with all students, not just those in finance.“Any college student should be interested on how well U.S. Capital markets work. Businesses use the markets to raise capital,” O’Hara said. “This allows them to expand and hire more people –– preferably Cornell graduates.”Ben Gitlin contributed reporting.
Original Author: Max Schindler