Following the method of Pesaran, Shin and Smith (1999), this study extends the results of Sun, Lin and Nieh (2007) to investigate risk diversification issue of individual corporate bonds in portfolios. Specifically our focus is on the decomposition of individual corporate yield spreads. It is also unique in the adoption of the robust econometric method of ARDL-based Pooled Mean Group estimation of panels of corporate bond data which yields results with solid statistical properties and rich economic implications for fixed income portfolio management. Overall the evidences provided by our study indicate that risk decomposition into systematic and specific components is crucial in a theoretical sense and useful practically as well. With the limited liquidity of corporate bond trading, our method proves to be an efficient one in extracting vital implications. More efficient and precise estimates would still require larger amount of data with better quality given the size and depth of the local fixed income market.