Identification of Insider Trading Using Network Numerical Models

Abstract

This article presents a network algorithm for identifying transactions which may constitute a violation of restricted periods, namely, making transactions in company shares by persons possessing inside information. The empirical research was performed on the basis of publicly available information on exchange trading, originating from the Warsaw Stock Exchange. The analysis is based on a numerical model which describes information spreading in a network with an information bottleneck. The applied method can confirm with high probability the use of inside information for carrying out unauthorized stock market transactions.

State-of-art review of information diffusion models and their impact on social network vulnerabilities

“…The identification of insider trading using network numerical models was presented in (Jakimowicz and Baklarz, 2016). In the CODA model, a change in the opinion (continuous variable) of nodes is revealed by observing the decisions (discrete variable) of neighboring nodes.…”

Impact of the Secondary Insider Trading on Financial Markets

“…This parameter was determined experimentally, depending on the volume, observed in the data, of large increases in the number of transactions that preceded stock exchange reports with information considered condential prior to such releases. This procedure was applied in previous studies involving 7 companies [6].…”