Since the inception of foreign exchange reform in 2005, China's campaigns for promoting its currency internationally are viewed as endeavours to cater to be commensurate with its increasingly vital position in global economy. This is a long journey to the end that is full of opportunities as well as challenges. Especially in recent years, Chinese currencies, both in onshore and offshore markets, have entered a volatile period in which many unexpected and surprising events happened, and with which the research on the topic of Chinese currency has been becoming heated. This research thesis aims to explore Chinese currency in various aspects in a broad context of ongoing renminbi internationalization, especially from the perspective of risk management. Unlike the traditional foreign exchange research in the context of macro-economics and international finance, we follow the route of quantitative research method and focus on data-driven statistical models. Specifically, in our research, we try to understand renminbi's behaviours from statistical and econometric perspectives, and use empirical approaches to modelling them. That leads to research contents and main contributions in the following three aspects.

The first piece of the thesis is to model the volatility of Chinese currency's foreign exchange rates. We at first go through the existing literatures in volatility modeling in financial assets. In order to model volatility, traditionally and still presently, academic scholars and industry practitioners use generalized autoregressive heteroskedastic models (GARCH) and its following variants which are known for their ability to account for stylized facts about time-dependent volatility such as clustering, fat tails and asymmetric effect. We pursue this line of work by modelling Chinese currency exchange rates and beyond that we propose a new volatility modelling method applied to Chinese currency which combines Markovian process and GARCH effect. Markov-regime switching GARCH model argues that volatility persistence comes not only from GARCH effect but also from switching between regimes with different levels of volatility. One of our results confirms that while onshore market is orderly in peace under central bank's guidance, offshore market is still in the high volatility regime.

After we understand the behaviour of the volatility of the Chinese currency foreign exchange rates, we further try to take a deeper look at Chinese currency by which we propose a novel empirical mode decomposition-based artificial neural network approach to research Chinese currencies with different investigated horizons in a progressive way. Empirical mode decomposition method (EMD) is first proposed for dealing with signal processing issues by decomposing original data sets into sub-components, which gives us an idea of considering multiple subset decomposed tasks instead of the whole one. Therefore, we combine the decomposition method EMD with the data analytical method called multi-layer neural network and find that the proposed method is superior to the pure data analytical one. Based on the results we get, we also test the decomposition model by carrying out a trading application and harness an positive annualized profit by using the proposed model.

Last but not least, we conduct an empirical research on hedging multiple portfolios against renminbi exchange rate risk, and the importance and necessity of hedging exchange rate risk in the backdrop of internationalising renminbi are affirmed and proved. With the internationalization of Chinese currency going forward and domestic institutions as well as individuals getting more exposed to international businesses, a major problem facing them is how to preserve the value of their assets and hedge the foreign exchange rate risk. In this part, we carry out an empirical analysis based on historical exchange rates data to show the importance of hedging the risk when investing overseas, and the results are examined from various perspectives ranging from Sharpe ratio to economic benefit.

In summary, Chinese currency renminbi is becoming a heated topic not only in industry but also in academic research. We would like to better understand the behaviour of renminbi and hope this awareness can gain us an edge in various decision makings. The rest of the thesis has the following structure. Chapter 2 gives literature review mainly on volatility modeling. Chapter 3 discusses the modeling of volatility of Chinese onshore and offshore currencies by using both traditional models such as standard and asymmetric GARCH models and the novel Markov-regime switching GARCH model. Chapter 4 puts forward a new EMD-based neural network decomposition model, and based on that we carry out a trading mechanism showing that the proposed model works well. Chapter 5 examines the necessity and importance of hedging Chinese currency's foreign exchange rate risk in the process of internationalization. Finally, Chapter 6 proposes some other research ideas together with future work and concludes the thesis.