Exploiting Big Data in the Financial Market - For the pursuit of better investment -

Public Relations Office        July 13, 2017

Katsuhiko Okada, Institute of Business and Accounting, Professor

Katsuhiko Okada, Institute of Business and Accounting, Professor

Joined Morgan Stanley New York as a derivative trader after receiving his MBA from Washington University’s John M. Olin School of Business in 1990. With 3 years of derivative trading experience in Morgan Stanley New York /Tokyo, and 4 years of managerial experience at the UBS arbitrage desk, he co-founded a hedge fund management firm (Halberdier Capital) in Singapore in 1997, which later grew to be one of the largest Japanfocused hedge funds in Asia. Katsuhiko sold his stake in 2001 and turned to academics. In 2005, he took up the position of assistant professor in Kwansei Gakuin University’s Institute of Business and Accounting. He became an associate professor in 2007 and a full professor in 2010. During this time, he received his PhD in Finance from Kobe University. His research interest is in behavioral finance. Katsuhiko serves as a director of academic associations such as ABEF and JFA and publishes influential papers regarding asset management in Japan.

From February 2015, Katsuhiko Okada, Professor of Finance at the Institute of Business and Accounting, has started providing investment advice to Mitsubishi UFJ Trust and Banking. His efforts are aimed at managing Japan’s first big data-driven investment fund. The Fund invests in Japanese equity and controls risk by monitoring market psychology. To accomplish this, his team, or their computer program to be precise, analyzes millions of news articles and professional comments every second. Here is a look at this cutting-edge project.

PC

“Buy the stock now!!” Unlike a standard mutual fund, trading orders do not come from fund managers. Instead, trading decisions are electronically disseminated from a computer program (artificial intelligence) and are distributed to fund managers at Mitsubishi UFJ Trust and Banking, who actually manage the fund. The logic of the algorithm was jointly developed by Prof. Okada’s laboratory and the Mitsubishi UFJ Trust Group. What the algorithm does, in essence, is quantify market sentiment and trade stocks accordingly. For instance, when investors’ opinion on TOYOTA is unanimously positive, the potential for a further rise upon good news is likely to be limited. Market sentiment is a good indicator to determine which stock is over-valued or under-valued. To calculate sentiment for all tradable stocks, they collect enormous amounts of textual data from the electronic news and social media. “We gather textual data from many sources, Twitter, blogs, Bloomberg news, message boards, to name a few, and we convert them into a machine-readable format. We then use our original classifier to detect whether the public mood is geared toward pessimism or optimism,” says Prof. Okada, who co-developed the program with his colleague Prof. Hamuro.

“Contrary to the traditional financial theory of Efficient Market Hypothesis, the stock market sometimes behaves irrationally and emotionally. Therefore, traders and fund managers should make decisions based on a wide range of information, including market sentiment. However, here comes the hard part: traders are also human and are biased when it comes to reading sentiment. Quantifying market sentiment under objective criteria helps, because it is an unbiased estimate of market psychology; and traders can rely on it.” Prof. Okada continues, “Our ultimate goal is to create virtual traders with artificial intelligence who trade stocks based on sentiment and fundamentals. They will be formidable traders.” The current algorithm will be updated constantly. Prof. Okada’s laboratory and the Mitsubishi UFJ Trust Investment Technology Institute (MTEC) will continue to work on the joint research project beyond 2015. Simultaneously, Magne-Max Capital Management Co., Ltd., an investment advisory firm founded by Prof. Okada, signed a contract, earlier this year, with Mitsubishi UFJ Trust and Banking as an investment advisory. In this unique setting, new ideas and innovative research will be instantaneously tested in the real world. “This sort of innovation in fund management was nurtured through collaboration between industry and academia,” says Toru Kigami, Executive Director at MTEC. Emphasizing the point, he adds, “Professor Okada and his team are key to the cutting-edge research which allows investor sentiment to be incorporated into our financial model, which we then take and put into actual practice.

Our success is due to our mutual collaboration.” Currently, Mitsubishi UFJ Trust and Banking are running a test fund investing approximately 1 billion yen of the company’s capital and, depending on the results of this trial, plan to expand and tailor it for customer use.