Elucidating Directed Statistical Dependencies: Investigating Global Financial Market Indices' Influence on Korean Short Selling Activities

Abstract

In previous studies, the short-selling market was a factor analyzed in a minor subject and mainly analyzed through macroeconomic factors or correlations. In this study, we tried to analyze the causal relationship with the short-selling volume of the Korean stock market by using Granger causality and transfer entropy, which is a different perspective compared to previous studies. The causal relationship between major global financial market indices and the prices of major community futures such as Bitcoin, gold, and silver and the amount of short selling in the Korean market was analyzed in the 20-day, 60-day, 120-day, and 240-day window. As a result, it was confirmed that very close linear and nonlinear directed statistical dependencies steadily occurred throughout the experimental period at the daily level. In particular, in the case of Granger-causality, it was confirmed that the causal relationship occurred the most from the price of crude oil futures in both the short term and the long term. In addition, the most frequent disambiguation of information from Saudi Arabia's market index, Tadawul All Share, can be interpreted as a result of having directed statistical dependencies from these oil-related indices due to Korea's high resource dependence. These results are significant because they provide a new perspective on macroeconomic market factors that researchers should consider when analyzing short selling.