Application Of Garch Model And Var Method In Foreign Exchange Portfolio

Investors always want to maximize expected returns in the investment process, and avoid risk as much as possible. Markowitz proposed portfolio theory, which answered the question that how to pursue the maximum yield at a given level of risk and how to minimize investment risk at a given return objective. He used the earning expectation and the capital earning covariance matrix to measure the risk, then established the portfolio model, using the model he got the portfolio efficient frontier. With the rapid development of investment theory, a large number of scholars have proposed new methods to measure investment risk, such as semi-variance, mean absolute deviation, semi-absolute deviation, value at risk (VaR). The VaR method has been widely used by financial institutions to measure the assessment risk, the Basel Committee recommended the VaR method as the standard method for banks to control financial risk, and other financial institutions also use the method.In foreign exchange market, many investors use technical analysis to trade, since investor’s ability is different, their trading strategies are often made based on experience in technical analysis and fundamental analysis, so their trading strategies often lack rigorous derivation. A large number of mathematical models and statistical analysis are used in quantitative investment to establish strict money management strategies to control risk, thus quantitative investment is becoming another mainstream financial investment method. This thesis will use the portfolio model to make trading strategy for multiple foreign exchange currency, containing EUR/USD, GBP/USD, USD/CHF, USD/JPY, AUD/USD, NZD/USD and USD/CAD. With the test of selected currency exchange rate, this thesis find the exchange rate yield has a peak and fat tail. For the above seven major currencies exchange rate, we chose GARCH (1,1) model to fit and predict the exchange rate, then obtain the earning expectation. When calculating the investment risk, GARCH model will be used to calculate the value of VaR. After getting earning expectation and investment risk, we establish single-period and multi-period foreign exchange trading model, which will be used to make trading strategy considering whether transaction cost exist or not. After verification, the model can maintain good profitability over time, which has a certain guiding effect for the establishment of quantitative trading model.The innovation of this thesis lies in revising GARCH model to predict the exchange rate, when making investment strategy, considering the trading cost, leverage and other factors, a multi-period model and a trading algorithm for foreign exchange quantitative trading has been proposed, which is consistent with the actual trading situation, so it has a certain practicality.