198 /302
供應鏈關係不確定性對公司信用風險影響之研究
198
Because a firm may have multiple customers or suppliers in a given year, we employ
Eqs. (1) and (2) to compute the FCRU and FSRU for each firm with multiple customers or
suppliers, respectively. In addition, we use a firm’s main customer or supplier CPSV or
KCICV as another proxy of FCRU or FSRU (denoted as M_CPSV/ M_KCICV).
(1)
(2)
Where
W
C
= CPS, C_BV (customer’s book value of assets), C_MV (customer’s
market value of assets), and C_Sales (customer’s net sales); W
S
= CIC, S_BV (supplier’s
book value of assets), S_MV (supplier’s market value of assets), and S_Sales (supplier’s
net sales)
We use bond yield spreads as the dependent variable and as the proxy of corporate
credit risk. A bond yield spread (SP) is defined as the yield difference between a corporate
bond yield and the yield of an equivalent-maturity Treasury bond. The SP data used in this
study are obtained from Datastream. Firm-specific control variables include leverage ratio
(LEV), equality volatility (VOL), R&D intensity (RD), industry concentration level
(HHI), and information asymmetry (ADJPIN) in addition to the probability of a
symmetric order flow shock (PSOS). Bond-feature control variables include the issued
amount (Lnamt), coupon rate (Coupon), life to final date (LFFL), bond age (Bage), and
Moody’s bond rating of each bond (Rating).
To execute empirical analyses, we estimate panel data regressions with firm- and
year-fixed effects by using 732 annual bond observations with customer identifications
and 1,733 annual bond observations with supplier identifications, respectively, during a
sampling period ranging from 2000 to 2008. We remove bonds with the following
characteristics: embedded options (such as convertible or callable bonds), floating-rate
coupons, security, government guarantees, and special clauses. In addition, the data for
bond data and other control variables are obtained from Datastream (e.g., bond data),
COMPUSTAT (e.g., financial data), CRSP (e.g., equity market data), and TAQ (e.g.,
intraday trading data) databases.