產業關聯風險、財務危機與公司債收益率價差

Abstract

計畫名稱：產業關聯風險、財務危機與公司債收益率價差 研究者：李漢星 經費來源：行政院國家科學委員會 近十來的金融危機已提升了信用事件預測與發展制定適切系統風險衡量指標的重 要性。先前的研究已指出資產關聯性在金融危機與熊市時增強 (Longin and Solnik, 2001; Chordia, Goyal, and Tong (2011))，雖然此現象於文獻已充記載，但對其背後所產生的原 因仍未有一致結論，金融危機時攀升的報酬相關性並無法被基本因子如總體經濟與公 司特徵所完整解釋。因此，學者們提出在機構處於財務危機條件下的系統風險的衡量 指標(Adrian and Brunnermeier, 2011; Acharya et al., 2010)。過去文獻亦已指出產業特性 可影響證券報酬與信用傳遞的方式(Hou and Robinson, 2006; Acharya, Bharath, and Srinivasan, 2007)。然而，僅有非常少的研究導入產業內的相關性風險進行公司債收益 率價差與信用分析。因此，在本研究中，我們嘗試研究產業內的相關性風險對公司債 收益率價差與破產預測影響。 本研究將首先進行理論與實證的文獻探討。我們將檢驗產業內關聯風險對公司債收 益率價差與違約預測的影響性。由於公司同時可能有發行超過一種以上的債券，我們 採用Petersen(2009)與Cameron, Gelbach and Miller (2011)等學者提出的方式，計算兩個 維度叢集的穩健標準誤以處裡時間序列以及異質變異問題。接下來，我們採用Duan et al. (2012) 的遠期違約強度模型，檢驗產業內關聯風險是否可在其他控制變數存在下仍 顯著影響公司違約風險。最後，我們將進行樣本外破產預測績效之分析，以檢驗加入 產業關聯風險因素可否增進破產預測之正確性

Title： Industry Correlation Risk, Financial Distress, and Corporate Bond Yield Spreads Principal Investigator： Han-Hsing Lee Sponsor：National Science Council

The financial crises in the last decade have raised the importance of the forecast of credit events and the need for devising appropriate measures of systematic risk. Previous empirical evidence documented that asset correlations increase during financial crisis and bear markets (Longin and Solnik, 2001; Chordia, Goyal, and Tong (2011)). Although it is well-documented in the literature that asset returns become more correlated during financial crisis, debate is still ongoing about the explanation of the increasing comovement in asset returns. The heightened correlation during crisis seems not be fully explained by the fundamentals such as macro factors or firm characteristics. As such, researchers have proposed measures for systemic risk conditional on institutions being under distress (Adrian and Brunnermeier, 2011; Acharya et al., 2010). Previous literature indicated that industry attributes can affect the security returns and how credit risks propagate (Hou and Robinson, 2006; Acharya, Bharath, and Srinivasan, 2007); however, there are few studies of corporate bond yield spreads and credit analyses incorporating the correlation risk among firms in an industry. Therefore, in this study, we attempt to approach this issue by investigating the impact of industry correlation risk on corporate bond yield spreads and default prediction. We will first review theoretical and empirical studies of industry correlation risk, asset correlation, corporate bond yield spreads, and default prediction. In our empirical test, we will examine the importance of industry correlation risk on corporate bond yield spreads and try to disentangle the effects of aggregate market risk and industry risk using several measures of industry risk. Given that each issuer may have more than one bond outstanding at any given point in time, we calculate two-dimensional cluster robust standard errors (Petersen, 2009; Cameron, Gelbach and Miller, 2011) to correct for time-series effect and heteroskedasticity. Next, we will use the forward intensity approach of Duan et al. (2012) to examine if industry correlation risk can affect default risk. This is the attempt to establish a linkage between industry correlation risk and credit risk. Finally, we will perform out-of-sample default prediction to examine if the inclusion of variables regarding market correlation risk can enhance the default prediction accuracy.