Page 42 - 臺大管理論叢第32卷第1期
P. 42

Extracting Liquidity Risk Factors by Credit Default Swap Quotation and Corporate Bond Yield:
               An Experimental Investigation



               Extracting Liquidity Risk Factors by Credit Default Swap
               Quotation and Corporate Bond Yield: An Experimental
               Investigation


               Chung-Ying Yeh, Department of Finance, National Chung-Hsing University
               Ren-Raw Chen, Graduate School of Business Administration, Fordham University, U.S.A.
               Bing-Huei Lin, Department of Finance, National Chung-Hsing University
               Shih-Kuo Yeh, Department of Finance, National Chung-Hsing University


                                                 1. Purpose



                    As widely known that corporate bonds are illiquid, yields of corporate bonds
               therefore contain significant premiums to compensate for insufficient liquidity. On the
               other hand, Credit Default Swap (CDS) products, ever since its introduction to the market
               in the mid 1990s, have been more actively traded. As a result, it is an interesting question

               whether or not their quotes include liquidity risk premium. Many scholars consider CDS
               spreads as a measure of pure credit risk, yet still others believe that the spreads affect
               prices and hence must be associated with liquidity or other risk factors.
                    This study tries to use a comprehensive GFI CDS dataset to estimate the liquidity

               premiums in bond yields and CDS quotes simultaneously using consistent econometrics
               methods. In particular, we measure the liquidity premium embedded in the bond yields to
               examine how CDS liquidity affects the yield spread of the corresponding bonds. Thus, we
               can provide the link between the liquidity premiums in both the CDS and the corporate

               bond markets.


                                                2. Approach



                    This paper employs two kinds of data. The first is CDS market quotations in the
               United States. Collected from the GFI CDS dataset, it covers all transactions of the period
               from July 2001 to October 2015. This dataset contains over ten million observations
               (4,135,728 bids, 3,762,940 asks, and 3,252,511 trades) across 2,677 reference entities.

               The other data is the corporate bond yield data collected from Trade Reporting and
               Compliance Engine (TRACE) matched with CDS market quotations. Moreover, we collect


                                                      34
   37   38   39   40   41   42   43   44   45   46   47