We are pleased to announce the winner of the Best Presentation Award that Entropy (ISSN: 1099-4300) sponsored at the 7th Workshop on Complexity in Economics and Finance, held on 19 October 2022 in Palma de Mallorca, Spain. Congratulations to Kiyoshi Kanazawa!

“Statistical analysis of a microscopic financial dataset for the long-range correlation of the order flow” by Kiyoshi Kanazawa

In the financial market, it is an established fact that order flow exhibits persistence: if you observe a buy (sell) order, you will likely observe a buy (sell) order even in the future. This phenomenon, called the long-range correlation (LRC), has been a topic under debate regarding its microscopic origin. One of the promising hypotheses in econophysics is the order-splitting hypothesis: in financial markets, many traders split their large metaorders into a series of small child orders. Because their order signs are kept the same during splitting, the market order sign has a (slight) predictability. This hypothesis was mathematically formulated by Lillo, Mike, and Farmer (LMF) in 2007. Interestingly, the LMF model predicts a quantitative relationship between a microscopic parameter (the power-law exponent of the metaorder-size distribution) and a macroscopic parameter (the power-law exponent for the order-sign autocorrelation function). In this talk, we present our statistical analysis to confirm the LMF prediction using a microscopic Tokyo Stock Exchange dataset. We develop a statistical method to measure the power-law exponents with less statistical bias, and we apply the method to confirm the LMF prediction. A long-standing problem in econophysics has been solved by our detailed data analysis of a microscopic financial dataset.