You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles
International Journal of Financial Studies
Download PDF
  • Article
  • Open Access

20 January 2025

How Has the Renminbi’s Role in Non-USD Currency Markets Evolved After COVID-19? An Analysis Based on Spillover Effects

,
,
,
and
1
School of Economics, Nagoya University of Commerce and Business, Aichi 470-0193, Japan
2
Graduate School of Asia-Pacific Studies, Waseda University, Tokyo 169-8555, Japan
3
Department of Economics, State University of New York at Binghamton, Binghamton, NY 13902-6000, USA
4
Faculty of International Media, Communication University of China, Beijing 100024, China
Int. J. Financial Stud.2025, 13(1), 12;https://doi.org/10.3390/ijfs13010012
Version Notes
Order Reprints

Abstract

Global uncertainty and the COVID-19 pandemic have significantly impacted the integration of emerging economies into global financial markets. Post-pandemic, the Federal Reserve’s interest rate hikes have drawn investor attention to relatively independent and stable currencies. This study investigates the sustained independence of the Renminbi by analyzing the spillover effects between the Renminbi and other major currencies in the context of the pandemic and USD interest rate hikes. By employing high-frequency data and cross-validating the results with low-frequency data transformed through Synchro Squeezing Wavelet Transform, we aimed to enhance the robustness of our findings. This analysis provides valuable insights for investors, highlighting the stability advantages of the Renminbi in the context of de-dollarization and global currency diversification.

1. Introduction

In recent years, global uncertainty has increasingly become a barrier to the integration of emerging economies into global financial markets. The outbreak of the COVID-19 pandemic has magnified spillover effects of leading currencies, with an impact on overall volatility spillovers approximately eight times that of the global financial crisis (Gunay, 2021; Hanif et al., 2023). In addition, according to the Global Financial Stability Report (2021), in contrast to the global financial crisis, the Financial Condition Index of emerging countries in the late stages of the COVID-19 pandemic is much lower than that of developed countries.
It is worth noting that the implementation of monetary policy in the United States needs to be analyzed over different periods due to different exogenous causes of uncertainty. The relationship between widening fiscal deficits and inflation in the United States has been confirmed and it has been found that expansionary monetary and fiscal policies did not lead to inflation in the 2007–2008 global financial crisis, but in contrast, the fiscal and monetary response to the COVID-19 pandemic led to inflation risk (Bordo & Levy, 2021). To control domestic inflation, since ending its four-year policy of zero interest rates in March 2022, the Federal Reserve has raised interest rates a total of 11 times over two years, with a cumulative increase of 525 basis points. This monetary policy shift has triggered significant global asset devaluation, including depreciation of non-USD (USD) currencies, and in particular, emerging market economies have faced increased debt servicing costs and heightened economic instability due to their reliance on dollar-denominated debt and transactions (Hoek et al., 2022), with the rise in interest rates prompted by inflationary pressures being even more disruptive to emerging financial markets. Moreover, exchange rate fluctuations, particularly in the context of Federal Reserve interest rate hikes, exert profound and multifaceted impacts on the current accounts of trade-surplus economies. The process is initiated by the Federal Reserve’s tightening of monetary policy, which reduces global dollar liquidity and escalates the cost of borrowing in USD. This development dampens global demand, leading to a subsequent reduction in the trade surpluses of non-U.S. economies.
Consequently, this accelerates the de-dollarization process, with several countries attempting to use other alternative currencies for trade settlements. Considering that the Renminbi is a proactive driver of de-dollarization and its international status has been steadily rising (Orăștean & Mărginean, 2023), and based on the analysis that the spillover effects of the Renminbi with other non-USD currencies gradually weakened during the COVID-19 pandemic and the independence of the Renminbi increased (Lu et al., 2023), we can observe the potential for the Renminbi to be favored by emerging countries as a safe asset. However, will this independence be sustained in the context of the twin eras of post-pandemic era and the USD interest rate hike? The answer to this question will have a significant impact on the process of Renminbi internationalization. More importantly, this study can provide empirical evidence to assess whether the independence of the RENMINBI is a temporary phenomenon triggered by the pandemic or a sustained trend likely to persist over the long term.
To build upon and extend their research, our paper contributes to the field in the following ways. (1) Extending the Data: We extended the dataset beyond the end of the pandemic, providing a more comprehensive analysis. (2) Enhancing Data Transformation: Unlike the CWT wavelet down-sampling method employed by Lu et al. (2023), we opted for the more effective Synchro Squeezing Wavelet Transform (SSWT). (3) Introducing Nonlinear Equations: In the VAR section, we introduce nonlinear equations to better align with the practical dynamics of currency exchange rate movements. (4) Bridging High-Frequency and Low-Frequency Differences: Our study aims to bridge the gap between high-frequency and low-frequency data by providing an in-depth analysis that elucidates the distinctions between the two. This approach is critical for a more comprehensive understanding of how currency dynamics unfold across different temporal scales. (5) We analyze the fragility of Renminbi independence in the face of significant fundamental fluctuations in the later stages of the COVID-19 pandemic. Through time regression analysis based on high-frequency data, empirical findings indicate an enhanced independence of the Renminbi. However, this independence is fragile due to the complex fluctuation relationships between the Renminbi and other non-USD currencies.
The structure of this paper is outlined as follows: Section 2 presents a review of the literature emphasizing the significance of employing high-frequency data. Section 3 details the data processing methodologies and the specifications of the econometric models. Section 4 focuses on the analysis of the empirical findings. Section 5 explores the application of Synchro-Squeezing Wavelet Transforms. Section 6 contrasts the analytical differences between high-frequency and low-frequency data. Finally, Section 7 concludes with a summary of the findings and provides policy recommendations and discusses the policy implications.

2. Literature Review

The literature review provides insights into the intensification of currency fluctuations caused by the pandemic and the significance of high-frequency data in analyzing volatility data. We found scholars unanimously acknowledge that the pandemic has exacerbated currency fluctuations, and high-frequency data offer advantages in elucidating volatility that low-frequency data lack.
Researchers have long recognized the importance of high-frequency data analysis in the field of financial data because financial time series exhibit significant heteroskedasticity, making volatility crucial in financial time series analysis (Zhou, 1996). However, estimating volatility from low-frequency time series is challenging. Yet, a considerable number of traditional studies on exchange rate spillover effects still rely on low-frequency data (Antonakakis & Kizys, 2015; Rajhans & Jain, 2015; Elsayed et al., 2022), primarily because low-frequency data are more readily available compared to high-frequency data. This situation has changed with the acceleration of digitization, especially during the pandemic, making high-frequency data more readily accessible. This has made analysis based on high-frequency data more feasible. If technological advancements have merely made high-frequency data analysis feasible, then the unique nature of the pandemic provides sufficient value for conducting high-frequency data analysis. Scholars believe that during the pandemic, market panic dominated the fundamentals of the economy, and herd behavior was prevalent in the markets (Aslam et al., 2020; Phiri, 2022). Grossmann et al. (2014) confirmed that high-frequency data have an advantage over low-frequency data in elucidating the sentiments behind forex trading and exacerbating exchange rate volatility. Therefore, it is crucial to utilize high-frequency data for currency exchange rate data research during the COVID-19 pandemic.
Furthermore, the Hildebrand et al. (2023) also highlights the significant impact of high-frequency exchange rate data on investment decisions and portfolio outcomes, which in turn confirms the limitations of traditional low-frequency data. They emphasize a major shift in the global economy since the onset of COVID-19, characterized by decelerated growth, exacerbated inflation, rising interest rates, and increased market volatility. In this context of uncertainty, it is imperative for investors to recognize the constraints of static asset allocation strategies in achieving anticipated returns. Consequently, dynamically adjusting portfolios are considered indispensable for effectively navigating the evolving market landscape. Incorporating these insights, this study utilizes high-frequency data to examine the relationship between the Renminbi and other non-USD currencies during different phases of the pandemic, further corroborating the findings with low-frequency data to highlight the spillover effects on currency dynamics.

3. Methodology

3.1. Data Wrangling

To examine the spillover effects between the offshore Renminbi and non-USD currencies during various stages of the COVID-19 pandemic, we adopted the time-period classification of the epidemic proposed by Lu et al. (2023) and extended it into five distinct phases: the initial outbreak in China, global dissemination, vaccine distribution efforts, emergence of the Omicron variant, and conclusion of the pandemic. These phases were chosen based on significant events that impacted the pandemic’s trajectory.
To ensure comprehensive analysis, specific sampling intervals were established for each phase. These intervals capture key periods that reflect the development and impact of the pandemic. A detailed description of the phases and their corresponding sampling intervals is provided in Table 1.
Table 1. Important events and sampling intervals of COVID-19.

3.2. Data Description

The dataset used in this research consists of high-frequency data collected at 30 min intervals and the most recent global currency payment rankings provided by SWIFT.1 Our analysis focuses on seven prominent non-USD currencies: offshore Renminbi (CNH), Swiss franc (CHF), Australian dollar (AUD), British pound (GBP), Japanese yen (JPY), European euro (EUR), and Canadian dollar (CAD).
Data were collected from 3 January 2020 to 5 May 2023, sourced from a reliable source, CHOICE. To ensure uniformity in data quality across different regions with varying official holidays and market closure times, we carefully selected time intervals with complete data coverage. This selection process resulted in a balanced dataset with 41,421 data points for each currency series.
The return of the exchange rate series was computed using the following equation:
R t = P t − P t − 1 P t − 1 × 100 %
where R t represents the return of the currency at time t and P t represents the closing exchange rate at time t . The non-USD currencies selected in this paper are all one-way exchange rates against the USD under the direct quote method, and the return of the USD index is used as a proxy variable for the return of the USD.
Table 2 and Figure 1 display descriptive information on the returns of each currency, where the data of Figure 1 are derived from descriptive statistics (Mean, Min., Max. and S.D.) summarizing the central tendency, variability, and range of the exchange rates across different periods.
Table 2. Descriptive statistics of the return series of all currencies.
Figure 1. Time series analysis based on descriptive statistics (Table 2).

3.3. Econometric Model Specifications

To effectively capture the significance and direction of volatility spillover effects between different financial time series, Lu et al. (2023) used the VAR-BEEK-GARCH model to examine the market’s spillover effects. Specifically, the vector autoregression (VAR) model was used to measure the mean spillover effects, and the BEKK-GARCH model was used to measure the volatility spillover effects. This paper adopts the above method and makes some adjustments in detail. The following figure demonstrates the framework of the model.

3.3.1. The Auxiliary Regression

Since the currencies we study all use the direct pricing method of USD, the return rate of the USD itself will affect the return rates of other currencies. We used auxiliary regression to eliminate this part of the influence (Shu, 2010; Shu et al., 2015) and replaced the residual series after regression with the non-USD currency return series.
R i t = C + R u s d x t + ∑ i = 1 8 D i t + r i t
where R i t represents the return of non-USD currencies and R u s d x t represents the return of the USD. The formula replaces e i t with r i t , indicating that the residual after regression is the independent rate of return of non-USD currencies. C is the constant term and Dummy variables D i t are used to examine the impact of the epidemic at different stages on the returns of non-USD currencies.

3.3.2. The Vector Autoregression Model

VAR (Vector Autoregression) is a model based on Ordinary Least Squares (OLS), making the inclusion of interaction terms acceptable. Various non-linear VAR models have emerged, such as LSTM VAR and XGBoost VAR, depending on the construction method (Ramey & Ramey, 1995; Dufour & Tessier, 1993). Recognizing that the mutual changes between different exchange rates are not unidirectional, this paper introduces nonlinear equations into the VAR model to enhance realism by incorporating non-linearities.
VAR(p) model before revision:
r t = C + ∑ i = 1 p A i r t − i + ε 0
where r t is a 7-vector of seven non-USD currencies’ return, p is the lag periods and ε 0 is the error term.
In the revised VAR model, r t is a 14-vector that contains the returns of non-USD currencies and their squares. By balancing a row, the following formula can be obtained:
C N H t = C + ∑ i = 1 p C N H a i C N H t − i − C N H b i 2 + ∑ i = 1 p J P Y a i J P Y t − i − J P Y b i 2 + ⋯ + ∑ i = 1 p C A D a i C A D t − i − C A D b i 2 + ε
Here, we take CNH as an example. CNH can be treated as a graph superimposed by several parabolas, while the other variables remain unchanged, and if one variable changes, the change trend of CNH’s return will have an opposite trend as the variable crosses the symmetry axis. As the model adjusts, so does the measure of mean spillovers. Since the slope of a point on a parabola represents the magnitude of the change, we measure mean spillover effects between currencies as the slope of the parabola. We can calculate the slopes of the points on the parabola where all the independent variables lie, and take the mean of the absolute values of the slopes to measure the mean spillover effects.

3.3.3. The BEKK-GARCH Model

This article adopts the BEKK-GARCH model proposed by Engle and Kroner (1995), which is set up as follows:
H t = C C ′ + ∑ k = 1 K ∑ i = 1 q A i k ε t − i ε ′ t − i A ′ i k + ∑ k = 1 K ∑ i = 1 p B i k H t − i B ′ i k
where H t represents the conditional covariance matrix. To simplify the formula, we set K = 1. Numerous academic studies have demonstrated that a BEKK-GARCH (1,1) model effectively captures the volatility spillover effects between different financial markets (e.g., Chng, 2009).
H t = h 11 , t h 12 , t h 21 , t h 22 , t ,   C = c 11 0 c 21 c 22 ,   A = a 11 a 12 a 21 a 22 ,   B = b 11 b 12 b 21 b 22
where C denotes the constant matrix, A is the coefficient matrix associated with the ARCH term, and B is the coefficient matrix of the GARCH term.
Specifically, coefficient a 12 captures the GARCH volatility spillover shock transmission from the return series r 1 t to r 2 t , while coefficient b 12 signifies the GARCH volatility spillover shock transmission from r 1 t to r 2 t . The Wald test is applied to evaluate these hypotheses. If the Wald test statistic exceeds the critical value, we reject the null hypothesis, indicating the presence of significant volatility spillover effects.

4. Analysis of Empirical Results at High-Frequency Data

4.1. Auxiliary Regression Analysis

The results (Table 3) reveal that the COVID-19 exerted a negative impact on offshore RENMINBI returns. While the pandemic’s various phases did not significantly affect the returns of other currencies, the patterns observed in the coefficients of the dummy variables remain insightful. The USD coefficients of Asian currencies’ return (CNH and JPY) are smaller than those of other regions. This is consistent with the judgment that currencies in the Asian region are relatively more independent of the US and European currency markets. In addition, the coefficients on the dummy variables for the return on the CNH are all negative, suggesting that the epidemic negatively affected the CNH returns in all periods of the epidemic. The large absolute value of the coefficients in the two phases of the epidemic outbreak in China (Period 2, −0.00004 *) and the large-scale closure of cities in China (Period 5-2, −0.00004 *) implies that these two phases had the most prominent negative impacts on CNH returns.
Table 3. Result of auxiliary regression of high-frequency data.

4.2. Analysis of Mean Spillover Effect

Given the nonlinear interactions between currency returns, a quadratic function is incorporated into the VAR model. After conducting a unit root test, we identified that the optimal lag order for the VAR model is 1. As detailed in Table 4, where CNH returns are treated as the dependent variable, there is a slight increase in the impact of the lagged returns of CNH and the Swiss franc on CNH returns. However, the absolute values of the coefficients for the remaining currencies, when related to the independent Renminbi return, show a decline. This trend implies that the mean spillover effects of other currencies on the independent Renminbi return diminished in the post-epidemic period. For instance, the mean slope of the AUD’s influence on the Renminbi shifted from 0.03 to 0.01, indicating a weakened impact of other international currencies on the Renminbi and a strengthening of the Renminbi’s independence. Notably, the coefficients representing the mean spillover effects of the JPY, CAD, and EUR returns on the CNH exhibit a substantial reduction, shifting from an order of magnitude of 0.1 to an order of magnitude of 0.01.
Table 4. Regression results of the mean equation of high-frequency data (CNH as the explained variable).
A similar trend is observed in the mean spillovers from the CNH to other non-USD currencies. As shown in Table 5, when the CNH is used as the independent variable, the absolute values of the coefficients representing the effects of the independent return of the CNH on the returns of other currencies declined. This indicates a weakening of the mean spillover effects of the CNH’s independent return on other currencies after the pandemic. For instance, the mean slope of the Renminbi’s impact on the Euro decreased from 0.015 to 0.003. This phenomenon suggests a diminished influence of the Renminbi on other international currencies, reflecting an increase in the Renminbi’s independence.
Table 5. Regression results of the mean equation of high-frequency data (CNH as the explanatory variable).
Although the mean spillover effects of CNH returns as the explanatory variable and the mean spillover effects of CNH returns as the explanatory variable both show a decline in the overall trend, there is a distinction in their values. Preceding the epidemic, the mean spillover effect coefficient of non-US international currency returns on CNH returns ranged from 0.01 to 0.1, whereas the mean spillover effect coefficient of CNH returns on non-US international currency returns was around 0.01. In a similar vein, post-epidemic, the mean spillover effect coefficient of CNH returns on CNH returns decreased to approximately 0.01, while the mean spillover effect coefficient of CNH returns on non-US international currency returns ranged from 0.001 to 0.01. This observation suggests an asymmetric relationship between the CNH and non-US international currencies, both before and after the epidemic, with the CNH being relatively weaker.
Collectively, the mean spillover relationship between CNH returns and the independent returns of the other six currencies underwent significant changes over the course of the three-year epidemic. When examined in phases, the fluctuations in the impact of various currency returns on the CNH during the epidemic can be classified into two distinct groups. In the first group, the mean spillover effect coefficients of four international currencies, specifically the CNH, CAD, EUR, and Swiss franc, with a time lag, displayed a consistent decreasing trend across most phases of the epidemic, with only a few instances of increase. And the second group exhibited a contrasting pattern: the influence of the JPY, GBP, and AUD predominantly showed an increasing trend in most phases, with only occasional phases of weakening. Conversely, the variations in the influence of CNH returns on non-U.S. currency returns exhibit a relatively uniform distribution. Moreover, following the conclusion of the epidemic, specifically in the sixth stage, there was a general decline in the mean spillover effect coefficient of CNH returns on non-U.S. currency returns, while the mean spillover effect coefficient of non-U.S. currency returns on CNH returns, on the other hand, generally increased.

4.3. Analysis of Volatility Spillover Effect

Table 6 presents the estimates of the overall volatility spillover effects between the CNH and other currencies in terms of independent returns.
Table 6. Estimates of volatility spillover effects of high-frequency data.
For the CNH and the JPY, the volatility spillovers exhibit a consistent pattern. In the first three phases, there is a peak in two-way volatility spillovers between the CNH and the JPY, marked by significant ARCH and GARCH volatility effects. During the fourth and fifth phases, the ARCH volatility transitions to a unidirectional spillover pattern before returning to its original level in the sixth phase. This suggests that the containment policies implemented during the later stages of the pandemic weakened the connection between the CNH and the JPY.
The CNH and the CAD exhibit a weakening trend in volatility spillovers. Initially, in the first phase, there is unidirectional volatility from the CNH to the CAD. In the second and third phases, this evolves into bidirectional volatility spillovers. By the sixth phase, the connection weakens further, transforming into a weak unidirectional volatility spillover from the CAD to the CNH. This indicates a relative decline in the influence of the CNH on the CAD.
The volatility spillovers between the CNH and the CHF show an increasing trend. In the first phase, only weak ARCH-type fluctuations are observed for the CHF concerning the CNH, while the CAD exhibits significant GARCH-type fluctuations. As the pandemic progresses, the volatility spillovers between the CNH and the CHF strengthen despite some fluctuations during the intermediate phases. By the post-pandemic stage, this robust volatility spillover effects remains, suggesting a heightened and sustained connection between the CNH and the CHF.
The CNH and the EUR exhibit noticeable strengthening in volatility spillovers. In the first phase, there is unidirectional ARCH-type volatility and bidirectional GARCH-type volatility between the CNH and the EUR. During the third and fourth phases, as the pandemic situation stabilizes, the two-way volatility spillovers reach their peak, characterized by significant ARCH and GARCH volatility. Although there is some reduction in the fifth phase, elevated volatility spillovers persist into the sixth phase. This indicates an increased influence of the CNH on the EUR.
The CNH and the GBP display a distinct pattern in volatility spillovers. Throughout all stages, the CNH maintains significant ARCH and GARCH volatility spillovers toward the GBP. However, the GBP’s volatility spillovers toward the CNH evolve from ARCH-type volatility in the first phase to GARCH-type volatility in the sixth phase, indicating a shift toward a more persistent volatility relationship over time.
In summary, the volatility spillover effects between the CNH and major currencies exhibit varying patterns during the different phases of the pandemic. While the connections with the JPY and the CAD weakened, the relationships with the CHF and the EUR strengthened, highlighting the evolving influence and independence of the CNH in the global currency market.
Volatility spillovers between the CNH and the AUD experienced a significant increase over the observed period. In the initial stage, the relationship was characterized by unidirectional ARCH-type volatility from the AUD concerning the CNH. As the pandemic evolved, the connection between the two currencies grew progressively stronger. By the sixth stage, this relationship developed into significant bidirectional ARCH- and GARCH-type volatility between the CNH and the AUD. This trend reflects an enhanced influence of the CNH on the AUD, underscoring the growing interconnectedness between these two currencies.
The WALD test reveals that all currency-independent returns exhibit bidirectional volatility spillovers with the CNH, but the significance varies. The CAD displays bidirectional arch-type volatility spillover effects only in stages I and IV. The CHF shows bidirectional arch-type volatility spillovers in Period 1 and GARCH-type volatility spillovers in Period 6, while the EUR demonstrates bidirectional arch-type volatility spillovers only in Periods 3 and 4, with varying degrees of significance. The GBP exhibits bidirectional ARCH volatility during specific periods from stages one to six. The AUD demonstrates two-way GARCH-type volatility spillovers only in stage I and stage II.
Combining the data from Table 6 and Table 7, CNH standalone returns can be broadly classified into three types of volatility spillovers with other non-U.S. currencies: The first category, characterized by substantial volatility spillovers and robust inter-currency connections, is predominantly represented by the CHF, the EUR, and the AUD. Before the epidemic, there was a one-way spillover effect between the CNH and these currencies, while after the epidemic concluded, this spillover effect became bidirectional.
Table 7. Joint test result of volatility spillover effects of high-frequency data.
The second category, featuring noteworthy volatility spillovers and weakened currency linkages, is mainly associated with the CAD. The volatility spillover between the independent returns of the CNH and the CAD transitioned from being significant before the epidemic to becoming insignificant after the epidemic.
In the last category, new characteristics of volatility spillovers emerged, and this category is primarily led by the GBP and the JPY. The volatility spillover effects of the CNH against the GBP and JPY did not undergo significant changes before and after the epidemic.

5. Analysis of Empirical Results Using Low-Frequency Data

Following the completion of the above analysis using 30 min high-frequency data, we applied wavelets to validate our previous findings. Wavelet transform, akin to the Fourier transform, serves as a data signal processing technique. However, unlike the Fourier transform, wavelets replace the infinite-length trigonometric basis with a finite-length wavelet basis that decays. This innovation addresses the Fourier transform’s limitation in capturing changes in the frequency of non-smooth signals over time. Generally, wavelet transform can be categorized into two main types: continuous wavelet transform (CWT) and discrete wavelet transform (DWT). The implementations of CWT and DWT differ in how they discretize the scale parameter, which is used to stretch or shrink copies of the basic wavelet and play a role in data decomposition.
However, in recent years, the Synchronous Squeezing Wavelet Transform (SSWT) has been developed as an alternative to empirical mode decomposition methods. SSWT employs a method called “synchronous squeezing” to adjust the width of the wavelet basis functions based on the local frequency content of the signal (Zhao et al., 2018; Marchi et al., 2023). This means that the width of the wavelet is adapted to match the scale of the features present in the signal at each point in time, allowing for more accurate representation of both high and low frequency components. The advantage of SSWT over CWT and DWT lies in its ability to achieve significantly higher time and frequency resolution through more flexible and reasonable wavelet components, which is important for capturing changes in the frequency of non-smooth signals over time (Chen et al., 2014; Wang et al., 2014). Hence, in this article, we employ SSWT as a means of frequency reduction.

5.1. Auxiliary Regression Analysis of Low-Frequency Data

In the low-frequency data auxiliary regression result (Table 8), the coefficient of the USD’s own return on the CNH return is −0.00417, which is smaller in absolute value compared to that of non-Asian currencies. This observation indicates that even when assessed over the long term, the returns of Asian currencies continue to exhibit significant independence. Additionally, the coefficients of the dummy variables for the CNH remain negative, with the absolute values of s2 and s52 being the most substantial. Once again, these results suggest that the epidemic has an overall negative impact on Renminbi returns, with the most pronounced influence occurring during the s2 and s52 periods. These findings align with the outcomes derived from the 30 min high-frequency data analysis.
Table 8. Result of auxiliary regression of low-frequency data.

5.2. Analysis of Mean Spillover Effect of Low-Frequency Data

Similar to the high-frequency analysis, we used the residuals from the regression to fit the nonlinear VAR model in the low-frequency part. After the unit root test, we found that the best fit coefficient of the VAR model is 4. Considering that the down sampling operation in this paper filters out the high-frequency data and retains the long-term trend, the lag coefficient of 4 implies that the wavelet transform is valid.
Due to space limitations, we present results only for data with a lag of 4. Table 9 reveals that when considering the CNH independent return as the explanatory variable, the absolute values of the coefficients representing the independent returns of the JPY, CAD, and AUD on the CNH independent return decrease. This suggests that the mean spillover effects of these currencies on Renminbi returns weakened after the epidemic. Conversely, the absolute values of the coefficients for the independent returns of three European currencies, namely the CHF, EUR, and GBP, on the CNH return increase, indicating that the mean spillover effects of European currencies on Renminbi returns strengthened after the epidemic.
Table 9. Regression results of the mean equation of low-frequency data (CNH as the explained variable).
In contrast, when using the CNH independent return as the explanatory variable (Table 10), the coefficients for the mean spillover effects of the CNH independent return on the independent returns of all non-U.S. currencies decrease to varying degrees. Specifically, the coefficients for the CNH independent return on the independent returns of the remaining non-U.S. currencies, excluding the JPY, decrease from the pre-epidemic order of magnitude 0.1 to an order of magnitude of 0.01, reflecting a significant reduction. This suggests a substantial weakening in the mean spillover effects of the Renminbi on these currencies after the epidemic.
Table 10. Regression results of the mean equation of low-frequency data (CNH as the explanatory variable).
In summary, the correlation between CNH returns and independent returns of non-U.S. currencies displays a weakening trend post-epidemic, consistent with the conclusions drawn in the high-frequency section. However, it is worth noting that specific currencies may exhibit variations between high-frequency and low-frequency analyses, which will be further examined later.

5.3. Analysis of Volatility Spillover Effect of Low-Frequency Data

Table 11 illustrates that the patterns observed in low-frequency data closely mirror those in high-frequency data, revealing a dichotomy in the relationship between the independent return of the CNH and the independent returns of other non-U.S. currencies—some growing stronger, while others weakening.
Table 11. Estimates of volatility spillover effects of low-frequency data.
The relationship between the CNH’s independent return and the independent returns of the AUD, CHF, EUR, and JPY becomes more prominent following the conclusion of the epidemic. In contrast, the volatility spillovers involving the independent returns of the CAD and the GBP weaken.
For example, the connection between the CNH and the JPY evolves from a unidirectional ARCH-type volatility spillover in the first stage to a bidirectional ARCH-type volatility spillover in the sixth stage, with persistent GARCH-type volatility spillovers in both directions. On the other hand, the relationship between the CNH and the CAD demonstrates the opposite trend, transitioning from a bidirectional ARCH-type volatility spillover in the first stage to insignificant volatility spillover effects in the sixth stage.
The Wald test results, as presented in Table 12, confirm bidirectional volatility spillover effects between the CNH and the JPY, CAD, and CHF in terms of independent returns. However, for the CNH and the EUR, as well as the GBP, the bidirectional volatility spillover effects disappear during the fourth stage of the first period. Similarly, the bidirectional volatility spillover effects between the CNH and the AUD vanish in the sixth stage of the first period.
Table 12. Joint test result of volatility spillover effects of low-frequency data.
These findings illustrate the dynamic nature of volatility spillover effects between the CNH and major currencies, highlighting how relationships strengthen or weaken across different phases of the epidemic and beyond.

6. Comparison of High-Frequency and Low-Frequency Analysis

The analysis of high-frequency and low-frequency data uncovers similar overarching trends but reveals notable differences when specific currencies are considered.
For mean spillover effects, the high-frequency data analysis shows a decline solely in the independent return coefficient of the CHF relative to the CNH. In contrast, the low-frequency data analysis indicates decreases in the independent return coefficients of the CHF, EUR, and GBP concerning the CNH. This highlights a broader weakening of mean spillover effects in the low-frequency context. Regarding volatility spillover effects, high-frequency data reveal a structural change in the relationship between the JPY and the GBP with the CNH during the epidemic, without a consistent trend of increase or decrease. On the other hand, low-frequency data demonstrate that the volatility relationship between the JPY and the CNH strengthens after the epidemic, while the connection between the GBP and the CNH weakens. These findings underscore the importance of analyzing spillover effects across different data frequencies to capture nuanced dynamics in currency relationships during and after significant economic disruptions.
As per Grossmann et al. (2014), high-frequency volatility is akin to the “nervousness” observed in foreign exchange trading, often indicating the uncertainty harbored by market participants regarding future price movements. This apprehension can result in more frequent trading and heightened price volatility, consequently exerting a more profound influence on financial markets and the broader economy. In contrast, low-frequency volatility primarily reflects long-term economic trends and cyclical changes, typically displaying greater stability. In essence, high-frequency data analytics return distinct results due to their incorporation of sentiment information that is unavailable in low-frequency data. Connected to the latter stage of the epidemic, the outbreak of the Russian–Ukrainian war and the Federal Reserve’s decision to increase interest rates triggered a series of events. Multiple bank failures within the U.S. led to panic among investors, prompting a temporary flow of funds to countries and regions in Europe. These sentiments were discernible in the high-frequency data, contributing to the disparities observed in the performance of the EUR, GBP, and JPY between high and low-frequency data analyses.

7. Conclusions

In this study, we investigated the impact of the 2019 coronavirus disease pandemic on the internationalization of the Renminbi. Utilizing the Synchronous Squeezing Wavelet Transform to analyze high-frequency data, we found that the pandemic resulted in changes in the relationship between the Renminbi and other currencies. Specifically, upon analyzing both high-frequency and low-frequency data, it becomes markedly evident that the overall correlation between the returns of Renminbi and those of non-USD currencies has diminished at the level of mean spillover effects. However, in terms of volatility spillover effects, divergences have transpired in the relationship between the Renminbi and other international currencies. The interconnectedness with certain currencies has strengthened, while relationships with others have attenuated. These findings reflect an enhancement in the independence of the Renminbi, while its independence remains highly fragile.
This paper, following the supplementation of post-pandemic data, asserts the ongoing presence of Renminbi independence. Such results suggest that the independence of the Renminbi can be attributed to additional factors. In light of the Hildebrand et al. (2023) projecting a high-risk, high-inflation (Board of Governors of the Federal Reserve System, 2024), high-debt financial environment post-2024, it is inferred that the underlying reasons for Renminbi independence lie in fundamental shifts within the global political and economic landscape. In detail, structural factors such as labor shortages, geopolitical conflicts, and the low-carbon transition constrain production, thereby rendering central banks unable to combat economic fluctuations through loose monetary policies as they did in the past. The urgency of global governments to restore their economies in the aftermath of the pandemic further exacerbates this issue. To make matters worse, sustained pressure on the banking sector in countries like the United States may potentially influence credit conditions in the foreseeable future, further exacerbating uncertainty surrounding economic prospects.
These research findings contribute to understanding the process of Renminbi internationalization and suggest the necessity of further studying the vulnerability of Renminbi independence in the face of significant fundamental fluctuations. For policymakers, the findings highlight the nuanced interaction between market movements and structural shifts. The Renminbi’s independence offers China’s monetary authority greater control over domestic conditions, reducing reliance on external policies. However, its fragile autonomy, influenced by complex currency fluctuations, necessitates vigilant monitoring and adaptive policies to ensure stability. For investors, the research reveals evolving currency interdependencies affecting portfolio allocation and risk management. Reduced spillover effects among non-USD currencies suggest diversification benefits and revised hedging strategies. The Renminbi’s resilience and vulnerabilities provide key insights into investment risks and opportunities in emerging markets.
It is important to note that the distinctive feature of the study lies in the utilization of high-frequency exchange rate data from both China and other developed countries. However, the drawback of these data is the lack of diversity. Specifically, our analysis is confined to exchange rate data from China as a single developing country and developed economies, which limits our ability to thoroughly explore the mechanisms through which new fundamentals affect the Renminbi, as well as the absence of a clear understanding of the performance of other developing countries under the new norm. Therefore, we will focus on two key directions in future work: firstly, delving into the mechanisms through which fundamentals influence the Renminbi by collecting additional high-frequency macro and microeconomic data; secondly, studying the performance of currencies from other emerging economies under the new fundamentals based on high-frequency exchange rate data. This will help uncover broader trends and insights, providing a deeper understanding of the global foreign exchange market.

Author Contributions

C.L.: Conceptualization; Funding acquisition; Supervision; Writing—review and editing; F.Y.: Data curation; Investigation; Writing—original draft; J.L.: Formal analysis; Methodology; Validation; Writing—review and editing; G.Z.: Project administration; Resources; Software; Visualization; Writing—original draft; L.L.: Conceptualization; Review and editing. All authors have read and agreed to the published version of the manuscript.

Funding

This paper is recognized and sponsored by the funding as shown below: Youth projects of National Social Science Foundation (China), including the “Study on the Global Renminbi Supply-Demand Mechanism from the Perspective of International Financial Public Goods” (Project No.: 19CGJ045).

Data Availability Statement

The data presented in this study are available on request from the corresponding author due to the data are part of an ongoing study.

Conflicts of Interest

The authors declare no conflicts of interest.

Note

1
In March 2020, the Society for Worldwide Interbank Financial Telecommunication (SWIFT) announced the top eight most traded currencies, namely, the US dollar, the European euro, the British pound, the Japanese yen, the RENMINBI, the Canadian dollar, the Australian dollar and the Swiss franc.

References

  1. Antonakakis, N., & Kizys, R. (2015). Dynamic spillovers between commodity and currency markets. International Review of Financial Analysis, 41, 303–319. [Google Scholar] [CrossRef]
  2. Aslam, F., Aziz, S., Nguyen, D. K., Mughal, K. S., & Khan, M. (2020). On the efficiency of foreign exchange markets in times of the COVID-19 pandemic. Technological Forecasting and Social Change, 161, 120261. [Google Scholar] [CrossRef] [PubMed]
  3. Board of Governors of the Federal Reserve System. (2024). Annual report of the board of governors of the federal reserve system: Monetary policy report. Board of Governors of the Federal Reserve System. [Google Scholar]
  4. Bordo, M. D., & Levy, M. D. (2021). Do enlarged fiscal deficits cause inflation? The historical record. Economic Affairs, 41(1), 59–83. [Google Scholar] [CrossRef]
  5. Chen, Y., Liu, T., Chen, X., Li, J., & Wang, E. (2014, October 26–31). Time-frequency analysis of seismic data using synchrosqueezing wavelet transform. SEG International Exposition and Annual Meeting (SEG-2014), Denver, CO, USA. [Google Scholar]
  6. Chng, M. (2009). Economic linkages across commodity futures: Hedging and trading implications. Journal of Banking & Finance, 33(5), 958–970. [Google Scholar]
  7. Dufour, J., & Tessier, D. (1993). On the relationship between impulse response analysis, innovation accounting and Granger causality. Economics Letters, 42, 327–333. [Google Scholar] [CrossRef]
  8. Elsayed, A. H., Gozgor, G., & Lau, C. K. M. (2022). Causality and dynamic spillovers among cryptocurrencies and currency markets. International Journal of Finance & Economics, 27, 2026–2040. [Google Scholar] [CrossRef]
  9. Engle, R., & Kroner, K. (1995). Multivariate simultaneous generalized ARCH. Econometric Theory, 11(1), 122–150. [Google Scholar] [CrossRef]
  10. Global Financial Stability Report. (2021, October). COVID-19, Crypto, and Climate Navigating Challenging Transitions. International Monetary Fund. Available online: https://www.imf.org/en/Publications/GFSR/Issues/2021/10/12/global-financial-stability-report-october-2021 (accessed on 30 November 2024).
  11. Grossmann, A., Love, I., & Orlov, A. G. (2014). The dynamics of exchange rate volatility: A panel VAR approach. Journal of International Financial Markets, Institutions and Money, 33, 1–27. [Google Scholar] [CrossRef]
  12. Gunay, S. (2021). Comparing COVID-19 with the GFC: A shockwave analysis of currency markets. Research in International Business and Finance, 56, 101377. [Google Scholar] [CrossRef]
  13. Hanif, W., Mensi, W., Alomari, M., & Andraz, J. M. (2023). Downside and upside risk spillovers between precious metals and currency markets: Evidence from before and during the COVID-19 crisis. Resources Policy, 81, 103350. [Google Scholar] [CrossRef]
  14. Hildebrand, P., Boivin, J., Li, W., Brazier, A., Kaminker, C., & Paul, V. (2023). Active approach: BlackRock 2024 global investment outlook. Available online: https://www.rl360adviser.com/news/2023/blackrock-global-investment-outlook-2024.htm (accessed on 18 November 2024).
  15. Hoek, J., Kamin, S., & Yoldas, E. (2022). Are higher US interest rates always bad news for emerging markets? Journal of International Economics, 137, 103585. [Google Scholar] [CrossRef]
  16. Lu, C., Li, J., Liu, L., & Yu, F. (2023). Spillover effect of the Renminbi and non-USD currencies after the COVID-19 pandemic: Evidence captured from 30-minute high frequency data. International Review of Economics & Finance, 84, 527–552. [Google Scholar] [CrossRef] [PubMed Central]
  17. Marchi, E., Cervetto, M., & Galarza, C. (2023). Adaptive synchrosqueezing wavelet transform for real-time applications. Digital Signal Processing, 140, 104133. [Google Scholar] [CrossRef]
  18. Orăștean, R., & Mărginean, S. C. (2023). Renminbi internationalization process: A quantitative literature review. International Journal of Financial Studies, 11(1), 15. [Google Scholar] [CrossRef]
  19. Phiri, A. (2022). Changing efficiency of BRICS currency markets during the COVID-19 pandemic. Economic Change and Restructuring, 55(3), 1673–1699. [Google Scholar] [CrossRef]
  20. Rajhans, R. K., & Jain, A. (2015). Volatility spillover in foreign exchange markets. Paradigm, 19(2), 137–151. [Google Scholar] [CrossRef]
  21. Ramey, G., & Ramey, V. A. (1995). Cross-country evidence on the link between volatility and growth. The American Economic Review, 85(5), 1138–1151. [Google Scholar]
  22. Shu, C. (2010). Impact of the Renminbi exchange rate on sian currencies. In Currency internationalization: Global experiences and implications for the Renminbi. Palgrave Macmillan UK. [Google Scholar]
  23. Shu, C., He, D., & Cheng, X. (2015). One currency, two markets: The Renminbi’s growing influence in Asia-Pacific. China Economic Review, 33, 163–178. [Google Scholar] [CrossRef]
  24. Wang, P., Gao, J., & Wang, Z. (2014). Time-frequency analysis of seismic data using synchrosqueezing transform. IEEE Geoscience and Remote Sensing Letters, 11(12), 2042–2044. [Google Scholar] [CrossRef]
  25. Zhao, Y., Cui, H., Huo, H., & Nie, Y. (2018). Application of synchrosqueezed wavelet transforms for extraction of the oscillatory parameters of subsynchronous oscillation in power systems. Energies, 11(6), 1525. [Google Scholar] [CrossRef]
  26. Zhou, B. (1996). High-frequency data and volatility in foreign-exchange rates. Journal of Business & Economic Statistics, 14(1), 45–52. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Download PDF
Factors Influencing FinTech Adoption Among Bank Customers in Palestine: An Extended Technology Acceptance Model Approach
Previous
Disclosure of Sustainability Information Under the Corporate Social Responsibility Directive: The Degree of Compliance of Portuguese Stock Index Companies
Next