Article Preview
TopIntroduction
Economic globalization is the inevitable result of contemporary social and economic development. Since China joined the World Trade Organization (WTO) in 2001, the connections between China’s economy and the world economy have been increasingly strengthened. Lahmiri (2017) proposed that the global economy’s integration, the economic interdependence among countries, and the monetary policies’ closer cooperation increase the exchange rate fluctuations’ complexity. After the exchange rate reform in July 2005, the RMB exchange rate continued to appreciate, and the fluctuation amplitude increased significantly. Therefore, Du et al. (2020), Nayak et al. (2016) and Gbatu et al. (2017) suggested that various microeconomic entities in China must improve their exchange rate risk management as soon as possible. From the exchange rate system reform in July 2015 to 2013, the RMB exchange rate maintained a long-term trend of one-way appreciation against the US dollar. As the US dollar has appreciated against other major currencies such as Euro, RMB also appreciated against Euro. Afterward, Nouira et al. (2019) put forward that the RMB exchange rate continued to fluctuate bilaterally and significantly, gradually leading to a market expectation of RMB exchange rate devaluation. In 2016, the RMB officially became a member of the basket of currencies of the International Monetary Fund. In such a complex and volatile foreign exchange market, any major international emergency may lead to great fluctuations in the RMB exchange rate, resulting in exchange rate risks and affecting China’s economic stability. Moreover, Lagat et al. (2016) stated that the RMB is not yet fully convertible in the balance of payments in the capital account, and the RMB exchange rate is still determined by market supply and demand conditions and exchange rate policy choices.
In the open economy, the exchange rate is an essential factor affecting a country’s economic development and a reliable basis for ensuring the international trade balance. As the relative price of two different countries’ currencies, the exchange rate has become a recognized and significant link to maintain international economic exchanges. Alagidede et al. (2017) indicated that the exchange rate fluctuations’ characteristics have always been scholars’ concern in international finance. The exchange rate prediction methods include basic analysis and time-sequence analysis currently. The basic analysis method predicts the long-term exchange rate fluctuations based on variables such as international exchange interest rates, gross domestic product (GDP), and balance of payments. It is not suitable for short-term exchange rate fluctuation prediction. Tehranchian et al. (2018) and Ouma et al. (2018) showed that the time-sequence analysis method predicts exchange rate fluctuation based on the information generated by the sequence changes over time, so it can also realize good prediction when applied to short-term fluctuation prediction. The Auto-Regressive Integrated Moving Average Model (ARIMA) model is adopted in the linear model for exchange rate fluctuation prediction. As a traditional linear model for time-sequence fitting, the ARIMA model can convert non-stationary sequences into stationary sequences for more precise fitting predictions. However, the focus of ARIMA model’s prediction lies in the sequence level, so linear expressions cannot accurately describe the complex exchange rate sequences increasing or declining nonlinearly, resulting in a limitation. Ehikioya et al. (2020) suggested that scholars in finance and economics had successively proposed the autoregressive conditional heteroscedasticity model and the generalized regression conditional heteroscedasticity model to solve the ARIMA model’s autocorrelation problem in the residual sequence. On this basis, some scholars proposed the Generalized Auto-Regressive Conditional Heteroskedasticity (GARCH) model. Adelakun et al. (2020), Zhang et al. (2018), Hossin et al. (2020) and Nguyen et al. (2017) held that the GARCH model performed better than the ARIMA model in exchange rate prediction, so it had become a significant research direction in modern econometrics.