The Effect Of State Ownership On Credit Bond Pricing And Information Production Of Credit Rating Agencies

Based on the stylized fact that state-owned enterprises(SOEs)and non-SOEs do not have equal status in the debt financing process,this thesis compares the differences in debt costs between SOEs and non-SOEs and their influencing factors using transaction data of three types of medium-and long-term credit bonds(enterprise bonds,exchange-traded corporate bonds,and medium-term notes)from2008 to 2020.This thesis first examines the direct effect of the state ownership on trade spreads,and thus identifies the investors’ differential risk attitudes towards SOE bonds and non-SOE bonds.Then,this thesis investigates how such differential risk attitudes affect the relationship between two pricing factors(liquidity,bond rating)and trade spreads.This thesis finds that SOEs’ debt costs are relatively lower than those of nonSOEs,which is consistent with the implicit government guarantee hypothesis.Implicit government guarantees can help SOEs reduce bankruptcy risk,hedge liquidity shocks and improve bond ratings,ultimately reducing their debt financing costs.However,it should not be overlooked that this expectation distorts the information production behavior of credit rating agencies(CRAs)to some extent,leading to the relative overestimation of SOE bond ratings and weakening the ratingspread sensitivity.This thesis first estimates the direct effect of the state ownership on the cost of debt.The quarterly trade spreads of SOE bonds are about 100 basis points lower than non-SOE bonds,which helps SOEs to reduce 30% of the debt cost.Further,the thesis finds that the negative effect of the state ownership on trade spreads is stronger for firms/bonds with higher default risks and in periods of higher economic policy uncertainty.When the expectation of implicit government guarantees is negatively shocked,trade spreads of SOE bonds with relatively higher ex-ante default risks rise more sharply.The above results suggest that bond trade spreads reflect investors’ expectations that SOEs may bear the implicit government guarantees,which makes investors demand a lower risk premium for SOE bonds,but the expectation is also fragile.Next,the thesis examines the time-varying characteristics of trade spreads of SOE and non-SOE bonds,and particularly focuses on the role of bond liquidity after2018Q2,during which the trade spreads of non-SOE bonds rise sharply.Based on a constructed bond-quarter level illiquidity measure,the thesis finds that market liquidity of non-SOE bonds declines more compared to SOE bonds after 2018Q2,and the positive liquidity effect(the impact of illiquidity on trade spreads)is also stronger for non-SOE bonds.The positive liquidity effect for non-SOE bonds is only confined to AA+ or below rated bonds after 2018Q2,while the liquidity effect is not significant or significantly negative for SOE bonds.By decomposing bond trade spreads,this thesis finds a higher co-movement between liquidity spreads of nonSOE bonds and their trade spreads after 2018Q2,but liquidity spreads of SOE bonds remain stable.The heterogeneity test suggests that large non-SOEs can hedge the bond market liquidity shocks.The above results suggest that non-SOEs suffered a more severe market liquidity shock after 2018Q2 compared to SOEs,and investors started to require risk compensation for bond illiquidity.Finally,the thesis examines SOE bond overrating and its impact on bond pricing from CRAs’ information production perspective.Using quarterly bond rating panel data,this thesis finds that after controlling for bond,firm,and macroeconomic characteristics,SOE bond ratings are still on average 0.5 to 0.7notches higher than non-SOE bonds,which reduces SOEs’ debt costs by about 25–35 basis points.The magnitude of SOE bonds overrating increases when the CRAs are facing more serious conflicts of interest and when issuers have higher default risks.By examining the rating-spread sensitivity,this thesis finds that both SOE and non-SOE bond ratings have information content.However,the rating-spread sensitivity of SOE bonds is significantly weaker than that of matched non-SOE bonds.This phenomenon only occurs when SOE bond ratings are relatively higher than matched non-SOE bond ratings.When SOE bonds are not overrated,their rating-spread sensitivity is not significantly different from that of matched non-SOE bonds.The heterogeneity tests indicate that the negative impact of SOE bonds overrating on the information content of ratings is stronger when the CRAs are facing more serious conflicts of interest.The above results are consistent with the hypothesis that implicit government guarantees may distort the information production behavior of CRAs.By reducing investors’ investment losses and CRAs’ reputation costs,government guarantees embedded in state ownership induce CRAs to inflate SOE bond ratings to collect more rating fees,and thus reduce the information content of SOE bond ratings.This thesis contributes to the literature in the following aspects.First,it adds to the literature on the debt costs of China’s enterprises.In particular,this thesis focuses on the effect of state ownership on the debt costs of China’s enterprises.Due to data availability,few studies have systematically quantified the debt costs for China’s enterprises.Using transaction data in the credit bonds,this thesis estimates the effect of state ownership on debt costs,which helps to understand the differences in debt costs between SOEs and non-SOEs.The results also confirm that the investors’ expectations about implicit government guarantees for SOEs are reflected in bond trade spreads.This thesis also finds cross-sectional heterogeneity in the value of implicit government guarantees,as evidenced by the fact that the value of implicit government guarantees is higher for SOEs with higher default risks in normal times.In addition,this thesis finds that the expectation is vulnerable.When the expectation is negatively shocked,SOE bonds with higher exante default risk are severely affected.Second,this thesis not only compares the cross-sectional differences in trade spreads between SOE bonds and non-SOE bonds,but also focuses on the differential changes of two.Specifically,this thesis examines the role of bond liquidity after2018Q2.The thesis also adds to the literature on the bond liquidity and liquidity effects in China’s credit bond market.Bond liquidity is an important pricing factor,but there are relatively few empirical studies on China’s bond market liquidity.The thesis finds that market liquidity of non-SOE bonds declines more compared to SOE bonds after 2018Q2,and the liquidity effects of non-SOE bonds are stronger than SOE bonds.Liquidity deterioration drives up the trade spreads of non-SOE bonds after 2018Q2.Third,the thesis also adds to the literature on rating inflation and its impact on the information content of bond ratings.The existing literature finds that SOE bond ratings are generally higher than non-SOE bond ratings,and this thesis investigates whether SOE bond ratings are overrated.The main reasons for rating inflation include conflicts of interest arising from the issuer paying model,rating-based regulation,and market competition.But rating inflation is the result of the collusion between issuers and CRAs,so the corporate ownership cannot be ignored.This paper finds that when the CRAs are facing more serious conflicts of interest,the magnitude of SOE bonds overrating is higher.Further,this thesis also finds that SOE bonds overrating has a negative impact on the information content of bond ratings.When SOE bond ratings are overestimated relative to the matched non-SOE bond ratings,the rating-spread sensitivity of SOE bonds is relatively weaker than that of non-SOE bonds.