Measuring dynamic dependency using time-varying copulas with extended parameters: Evidence from exchange rates data

This study proposes a novel approach that investigates the dynamic dependency among exchange rates by extending time-varying copulas' parameters following an autoregressive moving average (ARMA) process. The process consists of an autoregressive part that explains the effect of the previous parameters and a forcing variable that measures the dependence structure between marginal variables. We apply this model to the daily data of the exchange rates of five Asian countries with the strongest economies before and during the 2020 pandemic, namely CNY/USD, IDR/USD, INR/USD, JPY/USD, and KRW/USD. The ARIMA-GARCH model was used to model the exchange rates data and estimate the dynamic dependence using time-varying copulas with the extended parameters. The dynamic dependencies between Chinas and the four countries' exchange rates before and during the 2020 pandemic was evidenced. Moreover, India is the country whose exchange rate has been most strongly affected by the pandemic. Some of the highlights of the proposed approach are: This paper provides two algorithms to investigate the dynamic dependencies among exchange rates data during a crisis and forecast the data using time-varying copulas with the extended parameters. There are four extended time-varying copulas' parameters which can measure the dynamic dependencies between variables. The computation procedure is easy to implement. (C) 2021 The Author(s). Published by Elsevier B.V.