| Summary: | 碩士 === 銘傳大學 === 風險管理與保險學系碩士班 === 105 === Agricultural futures is the oldest commodity futures. During early stage of agricultural futures evolution, people had made agricultural products widely spread in the commodity market. And in late stage, most speculators began to focus on arbitrage of agricultural futures trading. But in the traditional trading strategies, most investors only consider how to make profit, instead of the risk which they need to pay attention. The purpose of this study is to explore the effectiveness of trading strategies which combine the value at risk (VaR) and technical indicators. Furthermore, the study approaches the suitability of the commodity futures and parameter sensitivity of indicators. This study uses daily closing price of futures (soybean、wheat、corn on CBOT and coffee on ICE), and the period of the date is from October 1, 2001 to September 30, 2016. The method of this study is using Historical Simulation Method to estimate VaR, and adding indicators MA、MACD and RSI to construct suitability trading strategies.
This study uses the Crystal Ball software analysis of probability density function, which found that the best fit distribution function of return including soy、 wheat and corn are the Lognormal function through Anderson-Darling Test. It is easy to observe that High Kurtosis and Fat Tail distribution function in Distribution of returns, which in line with general financial information character. This study uses R statistical software to analyze the suitability of the commodity futures and parameter sensitivity of indicators. The result of sensitivity analysis shows that the three targets can not apply the same parameter to make trading strategy. This is because each subject has its own unique features. According to the sensitivity of Initial amount of investment, the greater Initial investment amount is, the greater transactions and return rate are.
Finally, back testing of the trading strategies with indicators found that the strategy of single index is not suitable and the return rates are all negative in the three crops futures. However, the strategies of combining Value at risk and a single index indicator can produce good filter rule. According to these strategies, effect of MACD adding to VAR is the best, while RSI adding to VAR have the worst transaction effect. Nonetheless, increasing the initial amount can improve the trading results of strategy which is using RSI. In addition, the return of strategies which use more than two index indicators adding to VAR can not outcome the strategies which use only one index indicator combining with VAR. That is, it is too much stringent indicators that result in poor transaction number and return rate.
|
|---|
