Market anomalies and data persistence: The case of the day-of-the-week effect
Vol. 12, No 3, 2019
Alex Plastun
Sumy State University, Ukraine o.plastun@uabs.sumdu.edu.ua |
Market anomalies and data persistence: The case of the day-of-the-week effect |
Serhiy Kozmenko
University of Social Science, Lodz, Poland University of Customs and Finance, Dnipro, Ukraine kozmenko.uabs@gmail.com Viacheslav Plastun
Kyiv School of Economics, Ukraine plastun.v@gmail.com Hanna Filatova
Sumy State University, Ukraine hanna_filatova@ukr.net
|
Abstract. This paper investigates the degree of persistence in the financial markets’ data during different days of the week over the last twenty years. This allows taking a brand new look on the day-of-the-week effect and providing additional evidence against the efficient market hypothesis. The variety of the financial markets includes developed and emerging stock markets, FOREX, commodity and cryptocurrency markets. To measure the level of persistence the R/S analysis is used. The findings indicate that the level of persistence is different for different days of the week. This is inconsistent with the Efficient Market Hypothesis: data do not follow a random walk; and there can be indirect evidence in favor of the day-of-the-week effect. Conclusions on non-randomness of the data are important, because they allow choosing the best model to describe price dynamics so that to increase the predictive power of the existing models. Differences in the long-memory properties of the market data during different days of the week is an important finding that can lead to better understanding of the behavior of financial markets. High level of persistence implies data predictability, and therefore suggests that trend following technics can be applied to make profits from trading.
|
Received: January, 2019 1st Revision: April, 2019 Accepted: September, 2019 |
|
DOI: 10.14254/2071-8330.2019/12-3/10
|
|
JEL Classification: C22, G12 |
Keywords: market efficiency, anomaly, long memory, persistence, day of the week effect, R/S analysis |