Consequences of Mixed Sampling on Granger Causality – A Cautionary Note

Research output: Contribution to conferencePaperResearch

18 Downloads (Pure)

Abstract

This study analyses the effects of mixed sampling on Granger causality, where both temporal aggregation and systematic sampling filters are used in the same model. From a theoretical perspective, it derives a condition under which mixed sampling induces spurious Granger causality in a bivariate non-cointegrated setup. It finds that mixed sampling does not induce feedback causal distortion if causality runs from a stationary stock variable to either a stationary or a non-stationary flow variable. In all other cases, it induces a spurious bi-directional feedback Granger causality from a unidirectional causal relationship. An empirical exercise on the causality between the cryptocurrencies of Bitcoin and Ethereum illustrates the usefulness of the results
Original languageEnglish
Number of pages25
Publication statusUnpublished - 10 Jan 2023
Event2023 Asia Meeting of the Econometric Society, South Central and Western Asia, Mumbai, India - INDIAN INSTITUTE OF TECHNOLOGY BOMBAY (IITB), Mumbai, India
Duration: 10 Jan 202312 Jan 2023
https://www.econometricsociety.org/regional-activities/schedule/2023/01/10/2023-Asia-Meeting-of-the-Econometric-Society-South-Central-and-Western-Asia-Mumbai-India#:~:text=The%202023%20Asia%20Meeting%20of,K%20Narayanan%20(IIT%20Bombay). (2023 Asia Meeting of the Econometric Society, South Central and Western Asia, Mumbai, India)

Conference

Conference2023 Asia Meeting of the Econometric Society, South Central and Western Asia, Mumbai, India
Country/TerritoryIndia
CityMumbai
Period10/01/2312/01/23
OtherTime Series Econometrics, 20:00-21:45 Tuesday, 10 January, 2023
Internet address

Fingerprint

Dive into the research topics of 'Consequences of Mixed Sampling on Granger Causality – A Cautionary Note'. Together they form a unique fingerprint.

Cite this