Due to arbitrage limits imposed by credit or counterparty risk and margin requirements, CDS spreads may deviate from fundamentals driven by sentiment-driven investors who have an excessively bearish or bullish view of the market or the companies. I derive several systematic and firm-specific sentiment measures from index options and individual stock options, respectively, and I examine their impact on CDS spreads. Unlike previous studies that have focused on credit spreads from bond markets, I relate sentiment to CDS spreads because CDS spreads provide a valuable market-based assessment of credit conditions for hedging, speculation, and arbitrage and because CDS prices guide credit spreads from bond markets in terms of price. discovery (Blanco et al., 2005).
So when arbitrage is risky or limited, CDS spreads move in response to both fundamentals and sentiment. Alexander and Kaeck (2008) showed that the behavior of CDS spreads exhibits a pronounced regime-specific pattern. By examining the CDS indices, they found that in a turbulent market, CDS spreads are extremely sensitive to stock volatility.
In this paper, I discuss possible asymmetries in the responses of CDS spreads to investor sentiment and expect a more pronounced effect of sentiment during a turbulent period. The Arbitrage Limits Theorem illustrates that sentiment measures are more powerful in explaining CDS spreads during market shocks, especially for high-yield CDS spreads (typically for lower-rated firms). Our findings are in agreement with the study by Cao et al. 2010) who found that information from options markets can be useful in explaining CDS spreads.
Data and methodology
In the second step, this implied volatility is converted to form the fitted call or put prices, depending on whether. the strike price is below or above the current underlying price. Panel A provides the statistics for all CDS spreads and firm-specific variables, such as Skew, RelativeDemand, implied volatility and yield. This alleviates problems such as the presence of outliers, heterogeneity and non-normal errors. 2009) encouraged the use of quantile regression to create a robust and complete picture of the determinants of CDS spreads.
I use it to explore how CDS spreads behave in different amounts of credit risk distribution. In particular, the responses of CDS spreads to changes in explanatory variables may differ between the firms in the left tail (low credit. These coefficients are interpreted as different responses of CDS spreads in different quantiles to the same set of explanatory factors.
One of our purposes is to investigate whether the responses of CDS spreads to investor sentiment are regime dependent.
Empirical results
The negative coefficient for Skew indicates that the more negative the risk-neutral skew is perceived by investors, the higher the CDS spread. Models 4, 5 and 6 in Table 2 check the robustness of the relationship between CDS spreads and investor sentiment by controlling for several fundamental variables. The result of Model 4 shows that the relationship between sentiment proxies and CDS spreads remains significant even in the presence of the idiosyncratic fundamentals.
Overall, the relationship between sentiment proxies and CDS spreads is robust after controlling for market-wide and firm-level fundamentals, as shown in Model 6. To explore the sensitivity of CDS spreads to measures of sentiment in the samples, I ask whether this relationship is particularly pronounced at speculative companies with lower. As shown by the R-squares, our chosen sentiment proxies explain speculative firms' CDS spreads well.
In particular, the sensitivity of CDS spreads to investor sentiment proxies is significantly increased from investment group to speculative group. To provide a complete picture of the relationship between CDS spreads and investor sentiment, I conduct a quantile regression to examine how CDS spreads respond to sentiment proxies in different quantiles of credit risk spread. The conventional regression in Table 2 constrains the coefficients of sentiment proxies to be the same for all firms, implying that their impact on CDS spreads is similar for both high- and low-grade firms.
Our empirical results show that the relationship between CDS spreads and sentiment proxies is extremely strong in the upper tail of the distribution. Skew significance exists only for firms whose CDS spreads are located in the 75th and 90th quantiles. I reinvestigate the relationship between CDS spreads and this composite index for systematic sentiment and report the results in Table 5.
Therefore, the relationship between CDS spreads and the composite sentiment index is stronger in the upper tail of the distribution, similar to the result for each systematic proxy for sentiment (Table 4). The arbitrage limits hypothesis illustrates our findings that CDS spreads are more sensitive to sentiment during a volatile period.
Robust tests
The standard deviation in the turbulent regime is almost three times higher than in the calm regime. The significant relationship between investor sentiment and CDS spreads, as presented in the previous sections, implies an impact of aggregate errors in investors' beliefs about CDS spreads. To address the possibility that the results contain a rational assessment of credit risk, I decompose the sentiment measures by regressing each sentiment proxy on a set of rational predictors of default risk to obtain the remaining sentiment proxies, and then regress the CDS spreads on the remaining emotional proxies.
If the relationship is partially driven by the rational components, it will be significantly weaker in the presence of these rational control variables. Baker and Wurgles (2006) and Han (2008) used this approach to decompose proxies for investor sentiment when studying the price effect of sentiment in the stock and options markets. The rational variables I use include (1) the level of the risk-free rate (i.e., the five-year U.S. swap rate); 3) the demand for liquidity (i.e. the difference between the three-month USD swap rate and the government bond yield); and (4) the Baa-Aaa spread (i.e. the difference between the average Baa yield and the average Aaa yield.
This finding implies that the speculative sentiment measure contains rational updating in CDS spread valuation. For other sentiment measures, the magnitudes of their coefficient estimates hardly change in the presence of the rational component variables. The implied volatility smile is equivalent to negative skewness of the risk-neutral density of the stock return (Bollen and Whaley, 2004; Han, 2008).
Toft and Prucky (1997) proposed a skewness metric proportional to the slope of the implied volatility curve (or "smile") divided by the implied volatility of at-the-money options. BKM (2003) found a high correlation between the risk-neutral skewness and the slope of the implied volatility curve. A negative slope of the volatility smile, where the implied volatility of OTM puts is higher than that of at-the-money or in-the-money puts, corresponds to a negative skew in the risk-neutral density.
According to Bollen and Whaley (2004), the slope of the implied volatility curve is measured as the ratio of the average implied volatility for OTM puts (those with −0.375 <. Similar results can be found when examining the 50th and 90th quantiles, where the coefficient of the 90th quantile from the coefficient of the 50th quantile.
Conclusions
A higher CDS spread is associated with a steeper slope, while a lower CDS spread is associated with a flatter slope. After controlling for the underlying variables, the impact of investor sentiment remains significant, confirming that it affects CDS spreads. Furthermore, CDS spreads for lower-rated firms are more sensitive to the sentiment implied by the options.
A volatile CDS market hampers arbitrage activities due to higher transaction costs due to increased counterparty risk and margin requirements, especially for lower-rated reference entities. Therefore, option-implied sentiment affects CDS spreads, and its influence becomes even stronger when there are greater barriers to arbitrage in CDS markets.
Notes: This table reports the results of regression models examining the relationship between CDS spreads and investor sentiment proxies. The dependent variable is the CDS spreads of 375 sample firms with available stock option data. Sentiment readings and credit default swap spreads: Grouped by credit rating Sample Investment-grade Speculative-grade.
Notes: This table shows the sensitivity of CDS spreads to sentiment measures across samples with different credit ratings. The dependent variables are the CDS spreads for investment-grade and speculative-grade firms with stock option data available, respectively. Independent variables include firm-specific sentiment (Skew, relative demand for OTM put), systemic sentiment (individual and speculative sentiment, bullish consensus on S&P 500 futures from MarketVane, ratio of CBOE total equity to trading volume of placement and calls) and other basic. variables such as stock return, implied volatility, the return of the S&P 500 index and its implied volatility.
Notes: This table reports the relationship between CDS spreads and the composite index for systematic sentiment. The rational variables I used include (1) level of risk-free interest rate (i.e., the five-year US exchange rate);. (2) slope of the yield curve, or the "term spread" (ie the difference between the ten-year exchange rate and two-year exchange rate);.
(3) the demand for liquidity (i.e. the difference between the three-month USD swap rate and the government bond yield); and (4) the Baa-Aaa spread (i.e. the difference between the average Baa yield and the average Aaa yield of US corporate bonds). Spreads of credit default swaps and slope of the implied volatility curve: An alternative measure of risk-neutral skewness. Notes: The slope of the implied volatility curve is considered an alternative measure of risk-neutral skewness.
促进国家科学委员会资助计划的研究和开发成果的信息表 1999 年特别研究项目 研究成果摘要表 其他成果（无法量化的成果，如学术活动、奖项、重大国际合作和国际合作）研究成果的影响 电力和其他有助于工业技术发展的特殊有用项目等请填写文字描述。）。
结果项量化 内容和性质的名称或简要描述。
