Study on Interest Rate Risk of Taiwan Government Agency Bonds

• Main Authors: Chung-Chih Lai, 賴忠志
• Other Authors: Deng-Yuan Huan
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/01248762265848964146

碩士 === 輔仁大學 === 應用統計學研究所 === 93 === Abstract： Bond’s market grows rapidly since 1991 in Taiwan; therefore the transaction in second market is more popular, especially the cases of the government agency bonds which has the best liquidity in Taiwan’s market. Balance of risk and profit by bond’s market will easily appear form the decision that investors make. Thus, how to exactly keep abreast of bond’s risk is turning into a hot topic in investment market. The interest rate risk is greatly concerned by investors, because government agency bonds’ credit risk nearly doesn’t exist. How to precisely evaluate and predict price and value of the government agency bonds will need to be studied. In my thesis, I predict the daily price of Taiwan’s government agency bonds through neural network, then use genetic algorithms to evaluate suitable yield to maturity and duration set. The method is different from the other researches. The result in my thesis shows that neural network can easily predict the price of Taiwan’s government agency bonds without any previous hypothesis and mathematical model, and genetic algorithms can search for appropriate answer set of yield under the price and the date of delivery. I combine those powerful tools and offer investors a theoretical standard to make the decision with their experience.