Abstract
Due to the deepening of the internationalization of the renminbi (RMB), the gradual opening of the capital account and the further reform of the RMB pricing mechanism, the fluctuation of the RMB exchange rate has a markedly enhanced trend. Since 2014, the RMB against the US dollar has ended unilateral appreciation trend and gone into the two-way fluctuations. With the high speed of China’s economic development, the international influence of the RMB is increasing, and the international use of the RMB develops rapidly. In November 2015, International Monetary Fund (IMF) announced the RMB into the Special Drawing Right (SDR) currency basket. The RMB is now at a new stage of internationalization. The role of the exchange rate for the national economy can’t be ignored. Forecasting of exchange rate trend accurately and the analysis of exchange rate changes scientifically are of great significance in both theoretical and practical aspects.A new methodology for exchange rate forecasting is proposed in this thesis.
Firstly, from the perspective of China's exchange rate data analysis, this thesis studies systematically the lead-lag relationship between the offshore RMB and the onshore RMB under the influence of extreme events; from the perspective of the domestic economic situation, it analyzes the relationship between the major domestic economic variables and the exchange rate; from the perspective of the international economic situation, it analyzes the relationship between the RMB and the major foreign exchange assets in the process of the five times of exchange rate reforms; from the perspective of the public’s expectations, it also analyzes the correlation between the web search data, which is related to the public’s expectations, and the exchange rate.
Secondly, by the new comprehensive integrated approach, the exchange rate is finally weight integrated from four perspectives: the exchange rate data decomposition and integration, the domestic economic situation, the international economic situation and the public’s expectations.
The main innovations and features of this thesis are as follows:
(1) Theoretical research. This thesis analyzes the reasons of exchange rate fluctuations and develops a new methodology for exchange rate forecasting with four perspectives: the exchange rate data decomposition and integration, the domestic economic situation, the international economic situation and the public’s expectations.
(2) Basic data. In this thesis, the research data include not only exchange rate data and macroeconomic variable data, but also daily and monthly web search data, such as Baidu search volume index and Google search volume index, to forecast exchange rate. The research results show that there is a dynamic relationship between the web search data and the exchange rate. Modelling exchange rate based on the web search data can improve the forecasting accuracy.
(3) Empirical research. This thesis analyzes the dynamic relationship between exchange rate and four aspects, named the exchange rate data decomposition and integration, the domestic economic situation, the international economic situation and the public’s expectations, and forecasts the exchange rate based on those four aspects. In sum, the main innovations and features of the empirical research are as follows:
Firstly, a hybrid approach named EMD-Bry-Boschan method is proposed to study the lead-lag relationship between the offshore RMB and the onshore RMB under the influence of extreme events. Furthermore, this thesis analyzes the lead-lag relationship between CNH spot exchange rate and CNY spot exchange rate when the extreme events are caused by the market factors and/or the policy factors.
Secondly, a forecasting approach based on the exchange rate data analysis is proposed. It is the first time that the decomposition and integration approach named EEMD-LSSVR is utilized to forecast the exchange rate. EEMD algorithm is used to decompose the exchange rate to several different IMFs and a residual sequence, and LSSVR is employed to forecast those different IMFs and the residual sequence.
Thirdly, a forecasting approach based on the domestic economic situation is proposed. According to the Granger causality test and the correlation test, the thesis studies the correlation between the major domestic economic variables and the exchange rate, and develops VECM to forecast the exchange rate with three variables which are selected by the grey relational analysis (GRA).
Fourthly, a forecasting approach based on the international economic situation is proposed and combines the vine copula with the econometric forecasting model for the first time. The vine copula is used to study the dependencies between the RMB and the major foreign exchange assets, particularly during the five times of exchange rate reforms. And it examines three exchange rates, which are most relevant to RMB/USD, to develop VAR model based on the international economic situation.
Fifthly, a forecasting approach based on the public’s expectations is proposed. Daily and monthly web search data, such as Baidu search volume index and Google search volume index, which are related to the public’s expectations, can be used to forecast exchange rate. This thesis analyzes the dynamic relationship between the web search data and the exchange rate using the Granger causality test and the grey relational analysis, and develops VECM based on the public’s expectations.
Sixthly, a comprehensive integrated approach is proposed for exchange rate forecasting by using KELM to integrate the four aspects’ forecasting results, including the exchange rate data decomposition and integration, the domestic economic situation, the international economic situation and the public’s expectations. The empirical results show that the KELM integrated forecasting approach outperforms the benchmarks in level forecasting and directional forecasting.
| Date of Award | 10 Nov 2017 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Kin Keung LAI (Supervisor), Guangwu LIU (Co-supervisor) & Shouyang Wang (External Supervisor) |