Factor Analysis To Estimation Of Value-at-Risk Of Bond Portfolios

Bonds show appreciation for investors because of their stable returns, lower risk and other characters. Bond Assets usually are the most important part of the strategic investment portfolios, especially for institution investors with steady incomes such as Insurance Companies, Commercial Banks, Fund Companies and so on. But bonds' price is also impacted by various factors, and the interest rates are the most ones among them. The accelerating of our countries interest rates marketing and the Central Bank adjusting interest rates frequently increase bonds risk. Therefore, how to manage bonds' interest rate risk effectively becomes an important problem of bonds investors. In China, national debts have become the important part of financial market for over 20 years development. So this paper selects national debts as our studying objects. By using principal component analysis (CPA) method to study the zero coupon yield curve, we find that the movement of the curve yield can be simulated by two factors mostly, that is parallel shift and rotation. Then we quantitative them and get "synthetic parallel shift" and "synthetic rotation". On the basis we model Two Synthetic Factors (TSFs) to VaR estimation. We assume different bond portfolios, and estimate their VaR by using TSFs and Variance-Covariance methodology. We find that the VaR estimates by using TSFs are more accurate and it simples the computation of VaR, especially when the number of assets is large .So we can use it to estimate bond portfolios' risk and manage their risk. Although we use national debt data , the scholars demonstrate the fact by studying different countries' bond markets largely. So, TSFs to VaR Estimation Model can provide a kind of thought for other bonds' risk management.