Incorporating the RMB internationalization effect into its exchange rate volatility forecasting

https://doi.org/10.1016/j.najef.2019.101103Get rights and content

Abstract

Recently, the Chinese government has launched the renminbi (RMB) internationalization policy as an impetus to foster China’s global economic integration. The RMB internationalization effect on China’s economy and the RMB exchange rate has attracted massive attention in recent financial research. In this paper, we adopt a genetic programming (GP) method to generate new RMB exchange rate volatility forecasting models incorporating the RMB internationalization effect. Our models are proved to have significant accuracy improvement in predicting both RMB/US dollar and RMB/euro exchange rate volatilities, compared with standard GARCH volatility models, which are incapable of capturing the RMB internationalization effect. Furthermore, our models display salient practical implications for policy makers to formulate monetary policies and currency traders to design effective trading strategies.

Introduction

The renminbi (RMB) internationalization was an important financial development in China’s recent political framework along with the One Belt One Road (OBOR) initiative, which is the Silk Road Economic Belt and the 21st Century Maritime Silk Road policy (Funke, Shu, Cheng, & Eraslan, 2015). The OBOR initiative unfolds the blueprint for China’s economic integration into the world economy (Lien and Zhang, 2018, Du and Zhang, 2018). Promotion of the RMB’s external use, which is a crucial indicator of RMB internationalization, has become a requisite consideration against this political background.

There are ample benefits to a currency becoming internationalized. One example is the seigniorage owned by the issuing country, which indicates that the currency of the issuing country can be considered an asset without generating any interest to the holder (Blinder, 1996). Moreover, currency internationalization could increase its international impact in the global financial markets. As a result of RMB internationalization, RMB exchange rate volatility exerts significant influence on international financial markets (Qin, Zhang, & Zhang, 2018), where public information plays an important role on RMB exchange rate volatility (see Ho et al., 2017, Ho et al., 2018).

Nevertheless, there are also costs associated with currency internationalization. The monetary authority may lose control over domestic monetary conditions since a large amount of currency stock is held by foreigners. This might cause tremendous variations in the currency exchange rate once foreigners create large capital flows in the foreign exchange market (Tavlas, 1997). As a result, currency internationalization could play a vital role in affecting exchange rate fluctuation. Therefore, the RMB internationalization policy initiated by Chinese government in 2009 may encounter similar situations. It can be observed in Fig. 1 that the RMB/dollar exchange rate became much more volatile after 2016 compared with previous years. Regarding this newly initiated policy, a large number of discussions concerning the RMB internationalization effect on China’s economy and the RMB exchange rate have recently surged. Funke et al. (2015) and Ren, Chen, and Liu (2018) scrutinize the RMB internationalization effect on the onshore and offshore exchange rate pricing differential. They believe that loosening capital restrictions in Chinese financial markets could reduce the level and volatility of the pricing differential. Wang and Wang (2018) maintain that internationalization of the RMB could result in the RMB’s appreciation and that, more importantly, the policy exposes the RMB to international hot money, which might engender unexpected shocks for the RMB exchange rate.

Because the above studies have demonstrated that the RMB internationalization effect has impacts on RMB exchange rate volatility, the main purpose of this paper is to develop RMB exchange rate volatility forecasting models that can capture the RMB internationalization effect. However, existing volatility models, including GARCH models, may be incapable of integrating the RMB internationalization effect. Here, we adopt an artificial intelligence (AI) technology called genetic programming (GP) to generate relevant model forms. Ding, Zhang, and Duygun (2019) demonstrate that GP is a reliable method for creating new models. They use GP to incorporate the trading liquidity effect into the GARCH model, showing a considerable improvement in volatility forecasting compared with GARCH family models in forecasting both oil and shipping index volatilities. RMB internationalization is effectively an increase in liquidity and demand for RMB outside China. Therefore, the RMB internationalization effect could serve as a specific reflection of liquidity, and thus, our paper is consistent with previous work. In fact, the GP system can automatically identify the most relevant items in predicting volatility and combine those items into new models. In other words, if the RMB internationalization effect is irrelevant in predicting volatility, it will be absent in the final model representations. The system stores numerous model formations based on historical RMB exchange rate data. Then, the system recombines all model items according to the mechanism of accuracy evolution. The GP system then evaluates all possible model recombinations and produces the one most accurate forecasting model.

In addition, since we are investigating two RMB exchange rate time series, namely, the RMB exchange rate against the US dollar and RMB exchange rate against the euro, we also attempt to explore whether the volatility models for the two time series have the same model form. Unlike the traditional GARCH model, the GP system generates two different model formats for both the RMB/US dollar and RMB/euro time series, which exhibit quite different volatility patterns over time (see Fig. 1). We further demonstrate that two volatility forecasting models with the RMB internationalization effect retain superior forecasting performance compared with standard GARCH and GARCH-M models. Specifically, the accuracy of our new models in forecasting both RMB/dollar and RMB/euro improves by approximately 90% compared with that of a standard GARCH model in both in-sample and out-of-sample tests. On the other hand, the improvement rate of our models is approximately 80% compared with that of a GARCH-M model.

Therefore, it is clear that our model can provide more accurate exchange rate volatility forecasting than traditional methods of volatility forecasting can by including the RMB internationalization effect. Hence, our models hold elegant practical implications since exchange rate volatility forecasting is exceedingly helpful in various international economic activities. For example, Asteriou, Masatci, and Pílbeam (2016) reveal a considerably strong causal relationship from exchange rate volatility to import/export demand in less developed countries such as Indonesia and Mexico. Our models can thereby be useful for policy makers to formulate effective monetary policy to stabilize exchange rate fluctuations in order to boost import/export demand. Moreover, Della Corte, Ramadorai, and Sarno (2016) show that exchange rate volatility forecasting can be used to construct currency hedging and trading strategies, and their strategies achieve high profitability. As a result, our models could also be valuable for international traders in the foreign exchange market to design their currency hedging and trading strategies based on robust future volatility forecasting.

The remainder of the paper is organized as follows. Section 2 covers the data description, definitions of key variables including the RMB internationalization effect, and model setup for both traditional GARCH-type models and the new GP-based models. Section 3 introduces the GP system with its generated model forms for exchange rate volatility forecasting. In Section 4, we conduct model accuracy tests for both RMB/US dollar and RMB/euro exchange rate in-sample and out-of-sample data. Section 5 concludes the paper.

Section snippets

Data

The sample data that we collect are for the RMB exchange rate against the US dollar and euro on a daily basis from the CSMAR database. The sample covers the period from 1 January 2006 to 31 December 2018 (3186 daily total observations for each series). The RMB foreign exchange rate reform of China started in 2005. Before the reform, the RMB had been pegged to the US dollar, and the RMB exchange rate had little volatility. In 2005, People’s Bank of China announced the application of a managed

Preliminaries

In this section, we develop our GARCH model with the RMB internationalization effect based on the estimated variables in Section 2. For the specific model development, we adopt the genetic programming (GP) method from computer science. GP is an evolutionary computation (EC) technique inspired by biological processes (Banzhaf et al., 1998, Hirsh et al., 2000). There are two issues involved in developing a GARCH model with the RMB internationalization effect. First, the model format of the

Model accuracy tests

Section 3 accommodates issues of unknown model format as well as model accuracy evolution. Now that we have acquired the specific model formulation of volatility forecasting for both the RMB/dollar and the RMB/euro exchange rates, this section will address the second issue of model prediction accuracy. In particular, we compare our models with two GARCH models, namely, GARCH model and GARCH-M model, regarding both the RMB/dollar and RMB/euro exchange rate datasets for volatility forecasting

Conclusions

In conclusion, our paper employs an artificial intelligence technology called genetic programming to develop new exchange rate volatility forecasting models that take the RMB internationalization effect into account. Our models exhibit both theoretical and practical implications. From the GP-generated model outputs, we find that the RMB internationalization effect plays a key role in predicting volatilities in both RMB/US dollar and RMB/euro exchange rates. We further show that these two models

Acknowledgements

This paper is sponsored by the K.C.Wong Magna Fund at Ningbo University and is also supported by Ningbo and Chinese Academy of Social Science collaborative Grant No. (NZKT201701).

References (34)

  • E.P. Hong

    The autocorrelation structure for the garch-m process

    Economics Letters

    (1991)
  • K.-Y. Ho et al.

    Does news matter in china’s foreign exchange market? Chinese RMB volatility and public information arrivals

    International Review of Economics & Finance

    (2017)
  • K.-Y. Ho et al.

    Public information arrival, price discovery and dynamic correlations in the chinese renminbi markets

    The North American Journal of Economics and Finance

    (2018)
  • L.-G. Liu et al.

    Do external political pressures affect the renminbi exchange rate?

    Journal of International Money and Finance

    (2012)
  • S. Pong et al.

    Forecasting currency volatility: A comparison of implied volatilities and ar(fi)ma models

    Journal of Banking & Finance

    (2004)
  • J.C. Reboredo et al.

    Downside and upside risk spillovers between exchange rates and stock prices

    Journal of Banking and Finance

    (2016)
  • L. Yang

    A semiparametric garch model for foreign exchange volatility

    Journal of Econometrics

    (2006)
  • Cited by (14)

    • Portfolio constructions in cryptocurrency market: A CVaR-based deep reinforcement learning approach

      2023, Economic Modelling
      Citation Excerpt :

      Due to the rise of advanced technology, new techniques have been applied in the financial area. For example, genetic programming applications in volatility forecasting (Ding et al., 2019; Ding et al., 2020; Ding et al., 2021), and deep learning (DL) based approach has been applied in the portfolio optimization domain in the past few years (Bai et al., 2022). However, such approaches are undesirable because a large proportion of them are heavily based on price trend prediction models.

    • Does the regional proximity lead to exchange rate spillover?

      2022, Journal of International Financial Markets, Institutions and Money
      Citation Excerpt :

      Hassan and Tufte (1998) report similar findings for the Bangladeshi economy. Blue stream discusses the models to forecast the exchange rate volatility (Anjum and Malik, 2020; Ding et al., 2020; Leung et al., 2017; Taylor, 1987; Vilasuso, 2002) and the impact of exchange rate volatility on oil prices (Chen et al., 2022; Dai et al., 2020). As noted above, exchange rate volatility plays a very critical role in driving the economies and financial markets, thus, it is imperative to correctly quantify and forecast the exchange rate volatility (Anjum & Malik, 2020b).

    • Economic policy uncertainty and price pass-through effect of exchange rate in China

      2022, Pacific Basin Finance Journal
      Citation Excerpt :

      ERPT refers to the mechanism of converting the movement of exchange rates into the import price index (IPI) and then turn to the producer price index (PPI), as well as the consumer price index (CPI) (Phuc and Duc, 2019). RMB internationalization is one of the important financial development projects (Ding et al., 2020; Funke et al., 2015). Since evaluating the ERPT is important for Chinese monetary authorities (Ha et al., 2020).

    View all citing articles on Scopus
    View full text