Macroeconomic risks of supply chain counterparties and corporate bond yield spreads

Longstaff and Schwartz (1995) observe negative correlation between interest rates and credit spreads. Duffie et al. (2007) present that macroeconomic variables can improve the prediction power of reduced form credit models. Huang and Kong (2005) investigate the impacts of macroeconomic news announcements on credit spreads, and find that consumer confidence index, NAPM index and advanced retail sales have significant negative effects. Bonfim (2009) points out that economic growth is followed by an increase in new credit overdue, with a lag of at least 2 years. Besides, bank interest rates and government bond yields display a positive correlation with the cyclical component of new credit overdue, with a lag of 1 year. Tang and Yan (2010) show that default probability and credit spreads both decrease with the economic growth rate and increase with the volatility of the economic growth rate. They also demonstrate that credit spreads increase with cash flow volatility and decrease with firm-specific growth rate.

Besides, Caton and Goh (2003) also discuss whether a firm’s rivals are affected equally by its bond rating downgrades.

In addition, some studies discuss the relationship between trading policy and financial risk, such as Garci'a-Teruel and Marti'nez-Solano (2010). Garci'a-Teruel and Marti'nez-Solano (2010) show that supplier financing is a source of short-term finance and firms have a target level of accounts payable. The findings imply that a firm’s trading policies to supply chain counterparties affect its financial risk.

Tang and Yan (2010) employ the cash flow beta (defined as the coefficient of the macroeconomic indicator on a firm’s operating cash flow changes) measures the systematic exposure of firm-level cash flows. Firms with a high beta are more likely to perform well in an up market. In a down market, however, high correlation with the market is not desirable. In addition, Tsai et al. (2009) also show that bankruptcy prediction model with macroeconomic factors improves the prediction power of financial distress.

The shortcomings of the MS variable are that it is easily distorted by different business cycle stages (Tang and Yan 2010) and firms’ specific characteristics (e.g. cyclicality or counter-cyclicality (Mascarenhas and Aaker 1989; Korajczyka and Levy 2003; Brockman et al. 2010)). For example, though the operating performances of many firms are cyclicality dependent, some may be not or countercyclical.

Panel regression model with fixed effect is a very common model specification in the literature of bond yield spreads determinants, such as Campbell and Taksler (2003), Bhojraj and Sengupta (2003), Longstaff et al. (2005), Cremers et al. (2007), Chen et al. (2007), Wang and Zhang (2009), Tang and Yan (2010), Chen et al. (2011). In addition, the results of Hausman tests for all panel data regressions in this paper show that the fixed effect model is the more appropriate model specification. It is noted that the detailed results of Hausman tests for all panel data regressions are available upon request.